Coal Markets and Carbon Capture
Model Development and Climate Policy Applications
Dissertation
vorgelegt von
Dipl.-Ing. Roman Mendelevitch
von der Fakultät VII – Wirtschaft und Management der Technischen Universität Berlin Zur Erlangung des akademischen Grades
Doktor der Wirtschaftswissenschaften - Dr. rer. oec. – genehmigte Dissertation
Promotionsausschuss: Vorsitzender: Prof. Karsten Neuhoff, Ph.D.(TU Berlin) Gutachter: Prof. Sauleh Siddiqui, Ph.D. (Johns Hopkins University) Gutachterin: Dr. Franziska Holz (DIW Berlin) Gutachter: Prof. Christian von Hirschhausen (TU Berlin)
Tag der wissenschaftlichen Aussprache: 10. Oktober 2016
Berlin, 2016
Abstract The international consensus regarding the 1.5-2°C target in the COP21 Paris Agreement entails that most fossil fuel reserves must remain unburned. Currently, a majority of climate policies aiming at reducing fossil fuel consumption are directed toward the demand side. In the absence of a global carbon regime, these policies are prone to carbon leakage and other adverse effects. Supply-side climate policies present an alternative and more direct approach to reduce fossil fuel consumption by addressing its production. Here, coal, as both the most abundant and the most emission-intensive fuel, plays a pivotal role.
In the first part of this dissertation, I use a model of the international steam coal market (COALMOD- World) to examine the effects of different supply-side climate policies and the extent to which they can achieve desired reductions in coal consumption. The partial equilibrium model is designed to replicate global patterns of coal supply, demand, and international trade. It features endogenous investments in production, export, and transportation capacities in a multi-period framework, while allowing for substitution between imports and domestic production of steam coal. The first policy examines the introduction of taxes on steam coal exports and, alternatively, on steam coal production, based on the rationale of reduced consumption and improved terms-of-trade. To this end, COALMOD-World is extended to a two-level setting. Results show that while significant revenues can be generated through a tax, substantial reductions in coal consumption can only be induced if a large coalition of producers jointly introduces a tax. A second policy analysis investigates the effects of removing subsidies for steam coal production. While the policy has a small positive net welfare effect, prices do not increase by the magnitude required to drive down coal consumption. Another supply-side policy currently discussed is a moratorium on new coal mines. I compile a unique dataset of reserves in currently active mines. Using different estimates of reserves in China and India, I set up two scenarios. I find that the low estimate of reserves would result in a coal consumption pattern compatible with a 2°C target in the absence of the Carbon Capture, Transport and Storage (CCTS) technology.
CCTS technology has been used to justify prolonged use of coal for electricity generation. In the second part of this dissertation, I take a critical view on CCTS. Scenario results from the mixed- integer, multi-period, cost-minimizing network model CCTS-Mod suggest that - even in combination with CO2-enhanced oil recovery (CO2-EOR) - the technology will not play a major role in decarbonizing the European electricity system. It might be an alternative for industrial processes where CO2 can be captured at lower cost, especially in the iron and steel industry and in the cement industry. Coordination in developing transport and storage infrastructure is found to be crucial for realizing the associated economies of scale. Lastly, ELCO, an integrated modeling framework that is able to assess the implications of various regulatory approaches on the development of a future electricity mix with a detailed representation of CCTS is presented. The framework is applied to a case study of the UK Electricity Market Reform to illustrate the mechanisms and potential results attained from the model.
Keywords: Carbon capture, CCS, CCTS, climate change, coal, CO2-EOR, partial equilibrium modelling, policy analysis, reserves, resource markets, steam coal, strategic behavior, subsidy removal, supply-side climate policy, taxes I
Zusammenfassung Der bei COP21 in Paris erreichte Konsens eines 1.5°-2°C Ziels impliziert, dass große Teile der heute als Reserven verfügbaren fossilen Rohstoffe im Boden verbleiben müssen. Gegenwärtig zielt eine Vielzahl der Maßnahmen, die der Reduktion des Konsums fossiler Rohstoffe dienen soll, auf die
Nachfrageseite ab. Ohne ein verbindliches globales CO2-Budget, sind diese Politiken jedoch anfällig für Emissionsverlagerungen und andere unerwünschte Effekte. Angebotsseitige Maßnahmen stellen alternative Politikinstrumente dar, die durch Angebotsreduktion direkt eine Reduktion des Konsums fossiler Rohstoffe bewirken können. Dabei spielt gerade die Kohle als einerseits emissionsintensivster und anderseits reichlich vorhandener fossiler Rohstoff, eine zentrale Rolle.
Im ersten Teil dieser Dissertation nutze ich ein Modell des internationalen Kesselkohlemarkts (COALMOD-World), um die Effekte verschiedener angebotsseitiger Politikmaßnahmen auf die Nachfrage nach Kohle zu untersuchen. Dabei wird betrachtet, inwieweit diese dazu geeignet sind, angestrebte Reduktionen des Kohlekonsums zu erreichen. Das für die Untersuchung genutzte, partiale Gleichgewichtsmodell ist in der Lage sowohl das globale Angebot als auch die Nachfrage sowie internationalen Kesselkohlehandel zu simulieren. Über mehrere Zeitschritte hinweg umfasst es endogene Investitionen in Produktions-, Export- und Transportkapazitäten und ermöglicht die Substitution zwischen einheimischer Produktion und Importen. Das erste Politikszenario untersucht die Einführung von Export- bzw. Produktionssteuern auf Kesselkohle. Dieser Politik liegt die Idee zugrunde, dass dadurch einerseits der Konsum reduziert, andererseits, die Terms-of-Trade verbessert werden können. Zur Umsetzung des Szenarios wird COALMOD-World zu einem zweistufigen Setting erweitert. Ergebnisse zeigen, dass durch die Steuer zwar signifikante Erlöse erzielt werden können, jedoch für eine substanzielle Reduktion des Kohlekonsums eine große Koalition von Ländern, die gemeinsam eine Produktionssteuer erheben, notwendig ist. Als weitere Politikmaßnahme untersuche ich den Wegfall von Subventionen für die Kohleproduktion. Zwar zeigt die Maßnahme einen geringen positiven Effekt auf die Nettowohlfahrt, jedoch bewirkt sie keine erhebliche Steigerung des Kohlepreises. Diese wäre allerdings notwendig, um den Kohlekonsum nachhaltig zu reduzieren. Eine weitere angebotsseitige Politikmaßnahme, die aktuell diskutiert wird, ist die Einführung eines Moratoriums auf neue Kohleminen. Hierfür habe ich einen einzigartigen Datensatz zusammengestellt, der die Reserven in bereits bestehenden Minen zusammenfasst. Basierend auf verschiedenen Annahmen zu Reserven in China und Indien, untersuche ich zwei Szenarien. Die Ergebnisse deuten darauf hin, dass eine Beschränkung auf aktive Minen - unter der Annahme geringer Reserven in China und Indien - zu einer Nachfrage nach Kohle führen würde, die mit dem 2°C Ziel vereinbar wäre.
Dabei wird angenommen, dass die CO2-Abscheidungs-, Transport- und Speichertechnologie (engl. Carbon Capture, Transport, and Storage - CCTS) nicht zur Verfügung steht.
Im zweiten Teil dieser Dissertation wird die CCTS-Technologie, die lange Zeit dafür genutzt wurde, um neue Investitionen in Kohlekraftwerke zu rechtfertigen, kritisch beleuchtet. Ergebnisse des gemischtganzzahligen, kostenminimierenden Multiperiodenmodells CCTS-Mod zeigen, dass die Technologie keine große Rolle bei der Dekarbonisierung des europäischen Strommarkts spielen wird, selbst wenn das abgeschiedene CO2 zur tertiären Ölgewinnung (engl. CO2-enhanced oil recovery -
CO2-EOR) wertsteigernd eingesetzt wird. Die Technologie könnte jedoch eine Alternative für II
Emissionen aus industriellen Prozessen darstellen, bei denen das CO2 zu geringeren Kosten abgeschieden werden kann. Des Weiteren zeigen die Modellergebnisse, dass Koordination für das
Erzielen von Skaleneffekten bei Transport und Speicherung von CO2 eine zentrale Rolle spielt. Zuletzt wird mit ELCO ein integriertes Framework vorgestellt, welches für die Untersuchung unterschiedlicher regulatorischer Ansätze auf die Ausgestaltung eines zukünftigen Strommixes genutzt werden kann. Dabei ist die CCTS-Technologie mit ihren Prozessschritten im Detail abgebildet. Das Modell wird zur Untersuchung der UK Electricity Market Reform genutzt, um die abgebildeten Mechanismen und mögliche Modellergebnisse vorzuführen.
Schlüsselwörter: CO2-Abscheidung, CCS, CCTS, Klimawandel, Kohle, tertiäre Ölgewinnung mit
CO2, partiale Gleichgewichtsmodellierung, Politikanalyse, Reserven, Ressourcenmärkte, Kesselkohle, strategisches Verhalten, Abbau von Subventionen, angebotsseitige Klimapolitik, Steuern
III
Acknowledgements First of all, I would like to express my gratitude to my supervisor Prof. Christian von Hirschhausen. He sent me through a tough program of challenging and promoting. He did a great job in motivating me by providing access to incredible places: together, we have visited nuclear power plants, former and future nuclear waste disposals, have seen the strongest high voltage line in the world in Kazakhstan and met the avant-garde of energy modeling in Washington DC. I am thankful for these great experiences and opportunities and I am now eager to leave my own footsteps on the research field. I am also most thankful to my second supervisor Dr. Franziska Holz. From my very start at DIW Berlin she was always available to resolve the small and big issue of both, the everyday business and the bigger research plan. She was vital in guiding the research of my thesis and our discussions and meetings have held me on track. She was also pivotal for the fast conduct of my thesis allowing me to focus on my research and by keeping non-PhD related tasks at an acceptable level.
At this point, I also would like to gratefully acknowledge financial support by the DIW Graduate Center, and the Federal Ministry of Education and Research (BMBF). I also would like to thank my head of department, Prof. Dr. Claudia Kemfert, for providing the opportunity to work in a productive and resourceful environment in the department of Energy, Transport and Environment at DIW Berlin. My special thanks goes to Prof. Sauleh Siddiqui, who has shown me that even the most complex mathematical concept can be taught in an accessible and digestible way, and even more importantly, that it can be fun to work with it.
I would like to express my special thanks to my dear colleagues and companions at TU Berlin and DIW Berlin. Phillip Richter and Daniel Huppmann were great mentors to me who always had a good advice and an open ear. With Pao-Yu Oei working on the thesis has never felt like labor but like spending time with a good friend and doing something meaningful at the same time. During those years I have found a very productive co-author and even more important a very good friend with whom I share memorable moments both in the academic careers (our first conference in Vilnius) but also non-academic success be it on the beach volleyball cord or in the After-Work-Cup area. In both domains we have proven to be a good team and I hope we can further build on that both, in and outside academia. Special thanks also go to Jens Weibezahn, without whom I would have never started to work at TU Berlin, in the first place. Especially, in the last months of my dissertation Fabian Stöckl was a vital resource for discussions, proof-reading and divertissement for me. Many thanks for endless hours behind the desk and at the Späti. I would like to acknowledge the help of many other students who have helped me in preparing my thesis and who I very much enjoyed working with, at the forefront, my thanks goes to Kim Collins, Jan Ilsemann, Ivo Kafemann and Tim Scherwath.
I would like to thank my family for providing any possible type of support to me, raising me to be a critical mind but also hard-working and persistent person.
Last, but not least, great thanks and loving hugs go to Anni, who has always supported me even after a day when I again “have not managed to do anything” and stayed at work until deep at night, nevertheless. Thank you for keeping all the world’s troubles away from me so that I could focus on my research.
IV
Rechtliche Erklärung Hiermit versichere ich, dass ich die vorliegende Dissertation selbstständig und ohne unzulässige Hilfsmittel verfasst habe. Die verwendeten Quellen sind vollständig im Literaturverzeichnis angegeben. Die Arbeit wurde noch keiner Prüfungsbehörde in gleicher oder ähnlicher Form vorgelegt.
Berlin, 12. December 2016
Roman Mendelevitch
V Overview
Overview Chapter 1 Introduction ...... 1 Part A Coal Markets ...... 21 Chapter 2 The COALMOD-World Model: Coal Markets until 2040 ...... 22 Chapter 3 Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption? ...... 60 Chapter 4 Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters? ...... 78 Part B Carbon Capture, Transport, and Storage ...... 98 Chapter 5 Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe ...... 99
Chapter 6 European Scenarios of CO2 Infrastructure Investment until 2050 ...... 119 Chapter 7 The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO) ...... 141 Bibliography ...... 159 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4 ...... 185 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7 ...... 213
VI Detailed Table of Content
Detailed Table of Content Chapter 1 Introduction ...... 1 1.1 Motivation ...... 1 1.2 Rationalizing the new economics of steam coal ...... 3 1.2.1 Critical view on the Carbon, Capture, Transport and Storage technology ...... 3 1.2.2 A changing environment for coal ...... 6 1.2.3 Wedge of global institutional projections for coal demand ...... 7 1.2.4 Instruments of climate policy ...... 8 1.3 Mathematic modeling of policy incentives and games on international resource markets ...... 13 1.4 Outline of the dissertation ...... 15 1.4.1 Part A: Modeling the effects of climate policies in the world steam coal market ...... 15 1.4.2 Part B: The vision of CCTS as low-carbon solution for the electricity and industry sector ...... 16 1.4.3 Chapter origins and own contribution ...... 18 1.5 Concluding remarks and outlook for future research ...... 19 Part A Coal Markets ...... 21 Chapter 2 The COALMOD-World Model: Coal Markets until 2040 ...... 22 2.1 Introduction: A comprehensive coal market model is needed ...... 22 2.2 The international steam coal market ...... 24 2.2.1 Types of coal ...... 24 2.2.2 Coal markets ...... 24 2.3 Literature ...... 26 2.3.1 State of the international literature ...... 26 2.3.2 Development of the COALMOD model framework and publications ...... 27 2.4 The COALMOD-World model ...... 28 2.4.1 Overview...... 28 2.4.2 Model structure ...... 28 2.5 Model specification and input data ...... 33 2.5.1 Countries and nodes definition ...... 33 2.5.2 Production, costs, and reserves ...... 34 2.5.3 Land transport ...... 39 2.5.4 Export ports ...... 39 2.5.5 Freight rates ...... 42 2.5.6 Demand – Two possible scenarios ...... 45 2.6 Modeling results until 2040 ...... 46 2.6.1 Scenario assumptions: stagnating coal demand or climate policies with significant demand reduction ...... 46 2.6.2 Overview of results: stifle Asian “hunger” for coal ...... 48 2.6.3 Global trade results ...... 50 2.6.4 Price analysis ...... 55 2.7 Model limitations ...... 58 2.8 Conclusions ...... 59 Chapter 3 Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption? ...... 60
VII Detailed Table of Content
3.1 Introduction: Supply-side climate policies as an alternative route to achieve desired emission reductions ...... 60 3.2 Instruments of climate policy ...... 62 3.2.1 Demand-side policies ...... 63 3.2.2 Supply-side policies ...... 63 3.3 A production subsidy reform as a supply-side climate policy ...... 64 3.3.1 Definitions and data sources ...... 65 3.3.2 Findings from literature on coal production subsidies ...... 67 3.3.3 Current subsidies on coal production in selected countries ...... 67 3.3.4 Quantitative assessment: production subsidy reform ...... 68 3.4 A moratorium on new coal-mines as a supply-side climate policy ...... 69 3.4.1 Remaining coal reserves in operating mines ...... 71 3.4.2 Quantitative assessment: Mine Moratorium Scenario ...... 73 3.5 Conclusions ...... 76 Chapter 4 Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters? ...... 78 4.1 Introduction: A climate policy with a double dividend ...... 78 4.2 Model description and specification ...... 82 4.2.1 Upper level: Policy maker as Stackelberg-leader ...... 82 4.2.2 Lower level and data set specifications ...... 83 4.2.3 Solution algorithm ...... 83 4.3 Scenario definitions ...... 84 4.4 Discussion of results ...... 85 4.4.1 Export taxes on coal ...... 85 4.4.2 Production taxes on coal ...... 92 4.4.3 Comparison of results ...... 94 4.4.4 Qualification of results ...... 95 4.5 Conclusions ...... 96 Part B Carbon Capture, Transport, and Storage ...... 98 Chapter 5 Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe ...... 99 5.1 Introduction: the impact of the carbon capture, transport, and storage technology ...... 99
5.2 Modeling CO2-infrastructure ...... 101 5.2.1 Mathematical representation of CCTS-Mod ...... 102 5.3 Application of the model for Europe and used data...... 105
5.3.1 CO2 emission sources ...... 105
5.3.2 CO2 transport ...... 107
5.3.3 CO2 storage ...... 108 5.4 Different scenarios and their results analyzing political and geological uncertainties ...... 110
5.4.1 Reference scenario: certificate price increasing to 75 €/tCO2 in 2050...... 110
5.4.2 Certificate price increasing to 50 €/tCO2 in 2050 ...... 114
5.4.3 Certificate price increasing to 100 €/tCO2 in 2050 ...... 116 5.5 Conclusion: the future of a CCTS roll-out in Europe ...... 117
VIII Detailed Table of Content
Chapter 6 European Scenarios of CO2 Infrastructure Investment until 2050 ...... 119
6.1 Introduction: CO2-Enhanced oil recovery keeps the mirage alive ...... 119 6.2 Model, data, and assumptions ...... 121 6.2.1 The model CCTS-Mod ...... 121 6.2.2 European data set ...... 123 6.2.3 Mathematical formulation ...... 125 6.2.4 Assumptions for all scenarios ...... 126 6.3 Results of the European-wide scenario analysis ...... 127 6.3.1 EU_40% scenario ...... 127 6.3.2 EU_80% scenario ...... 127 6.3.3 Sensitivity to investment and variable costs...... 129 6.3.4 Summary of the European-wide scenarios ...... 130
6.4 Regional focus: CO2-enhanced oil recovery options in the North Sea and the Role of regional Cooperation ...... 131
6.4.1 The role of CO2 reuse for CCTS ...... 131
6.4.2 CO2-EOR resources in the North Sea ...... 132
6.4.3 Costs and revenue of CO2-EOR ...... 132
6.4.4 Regional scenario: NorthSea_40% scenario with CO2-EOR option ...... 134
6.4.5 Regional scenario: NorthSea_80% scenario with CO2-EOR option ...... 135
6.4.6 Regional scenario: DNNU_80% scenario focusing on CO2-EOR in DK, NL, NO and UK ...... 137
6.5 Conclusion: the importance of CO2-EOR for a European CCTS roll-out ...... 139 Chapter 7 The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO) ...... 141
7.1 Introduction: a review of state of the art electricity and CO2 modeling approaches...... 141 7.2 Mathematical representation of the ELCO model ...... 144 7.2.1 Notations of the model ...... 145 7.2.2 The electricity sector ...... 147 7.2.3 The electricity transportation utility ...... 149 7.2.4 The industry sector ...... 150
7.2.5 The CO2 transportation utility ...... 151 7.2.6 The storage sector ...... 151 7.2.7 Market clearing conditions across all sectors...... 152 7.3 Case study: the UK Electricity Market Reform ...... 152 7.3.1 Describing the instruments: Contracts for Differences, Carbon Price Floor, and Emissions Performance Standard ...... 153 7.3.2 Data input ...... 155 7.3.3 Case study results ...... 156
7.4 Conclusion: findings of an integrated electricity-CO2 modeling approach ...... 157 Bibliography ...... 159 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4 ...... 185 A.1 Model Details: mathematical formulation, node structure, data, and results ...... 185 A.1.1 Set, parameters, and variables ...... 185
IX Detailed Table of Content
A.1.2 Producer formulation ...... 187 A.1.3 Exporter formulation ...... 189 A.1.4 Final demand formulation ...... 191 A.2 Model extension: Policy maker as Stackelberg-leader at the upper level ...... 191 A.2.1 Export tax ...... 191 A.2.2 Production tax ...... 192 A.3 MPEC solution algorithm ...... 198 A.3.1 First step: GAMS NLPEC solver ...... 198 A.3.2 Second step: disjunctive constraint reformulation ...... 199 A.4 Global trade flows in 2030 by case ...... 200 A.5 Sensitivity analysis: discount rate and tax growth rate ...... 201 A.6 Coalition joint by the USA ...... 201 A.7 Country-by-country assessment of coal production subsidies ...... 203 A.8 Country-by-country assessment of coal reserves in operating mines ...... 206 A.9 Further results of scenario with high estimate of reserves in operating mines ..... 209 A.10 Further details and results of the M&E scenario ...... 210 A.11 Conversion between reserves and resources data from McGlade and Ekins (2015) and COALMOD-World ...... 211 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7 ...... 213 B.1 CCTS-Mod: additional data and results ...... 213 B.2 Combined Electricity and CCTS Investment and Dispatch Model (ELCO): Karush-Kuhn-Tucker conditions ...... 217 B.2.1 The electricity sector ...... 217 B.2.2 Shared environmental constraints for the electricity sector ...... 218 B.2.3 The electricity transportation utility ...... 219 B.2.4 The industry sector ...... 219
B.2.5 The CO2 transportation utility ...... 220
B.2.6 The CO2 storage sector ...... 221 B.2.7 Market clearing conditions across all sectors...... 222
X List of Tables
List of Tables Table 1.1: Examples for different types of supply-side and demand-side climate policies...... 9
Table 1.2: Chapter origins ...... 18
Table 2.1: Major steam coal producers and consumers in 2014...... 25
Table 2.2: Assumed production capacity expansion limitations per five-year period (in Mtpa)...... 37
Table 2.3: Energy content of coal by production node...... 38
Table 2.4: Assumed export capacity expansion limitations per five-year period (in Mtpa)...... 41
Table 2.5: Freight rates for selected routes (in USD/t)...... 44
Table 2.6: Reference consumption in 2020, 2030, 2040, by IEA region from New Policies and 450ppm Scenario, and extrapolation for 2050 (in % of 2013 consumption)...... 46
Table 2.7: Share and rank in international trade flows of major exporters in both scenarios and over time...... 54
Table 2.8: Export capacity and production capacity: results of 2°C scenario compared to Stagnation scenario (in Mtpa)...... 55
Table 3.1: Total subsidy in 2013/2014 and subsidy per unit of production and by region for main coal producing countries...... 68
Table 3.2: Effect of subsidy removal on producer, exporter, and consumer surplus...... 69
Table 3.3: Estimates of resources and reserves from literature, and own estimates on reserves in operating mines...... 72
Table 3.4: Cumulative production in the reference case and in the Mine Moratorium scenario in Gt. . 73
Table 5.1: Investment costs for capture facilities (in €/tCO2pa). . (dimensioning of capturing sytem) 106
Table 5.2: Variable costs for CO2 capture (in €/tCO2)...... 106
Table 5.3: Investment cost by pipeline diameter and respective annual transport capacity...... 108
Table 5.4: Site development, drilling, surface facilities and monitoring investment cost for a given
annual CO2 injection rate per well...... 109
Table 5.5: Key scenario assumptions...... 110
Table 5.6: Overview of scenario results ...... 117
Table 6.1: CO2 certificate price path in the different scenarios...... 126
Table 6.2: List of scenario assumptions...... 127
Table 6.3: Input parameters for sensitivity analysis, and reference values for comparison...... 129
Table 6.4: Summary of the European-wide results...... 131
Table 6.5: CAPEX and OPEX cost components for CO2-EOR installation...... 133
XI List of Tables
Table 6.6: Cost and revenue items for the deployment of combined CCTS and CO2-EOR technology...... 134
Table 6.7: Average investment costs in CO2 transport and CO2 storage per MtCO2 per year, comparing the NorthSea_80% and DNNU_80% scenarios...... 138
Table 6.8: Summary of regional results...... 139
Table 7.1: List of sets of the ELCO model...... 145
Table 7.2: List of parameters of the ELCO model...... 145
Table 7.3: List of variables of the ELCO Model ...... 146
Table 7.4: List of dual variables of the ELCO Model ...... 147
Table A.1: List of sets in the COALMOD-World model...... 185
Table A.2: List of parameters in the COALMOD-World model...... 185
Table A.3: List of variables in the COALMOD-World model...... 186
Table A.4: Nodes of COALMOD-World...... 193
Table A.5: World Energy Outlook demand projections for coal for power generation in the scenarios (Mtoe)...... 194
Table A.6: Various input parameters for COALMOD-World production nodes...... 194
Table A.7: Results of COALMOD-World: consumption, domestic supply, and imports by consuming country and scenario in 2010, 2020, 2030, and 2040...... 195
Table A.8: Results of COALMOD-World: domestic supply and exports by producing country and scenario in 2010, 2020, 2030, and 2040...... 195
Table A.9: Trade flows in COALMOD-World (in Mtpa)...... 196
Table A.10: Comparison of key statistics across scenarios for at the tax-revenue maximizing level, and at 10 USD/t as the initial tax rate...... 197
Table A.11: Cumulative production in reference case and scenario with high estimate of reserves in operating mines (in Mt)...... 209
Table A.12: Cumulative production in reference case and M&E scenario (in Mt)...... 210
Table A.13: Production capacity and reserves in COALMOD-World dataset and M&E scenario...... 212
Table B.1: Definition of indices, parameters, and variables of CCTS-Mod ...... 213
Table B.2: Estimated CO2 storage potential ...... 213
Table B.3: Summary of scenarios results and CO2 price assumptions for Chapter 6 ...... 216
XII List of Figures
List of Figures Figure 1.1: The process chain of carbon capture, transport, and storage...... 4
Figure 1.2: Projected coal demand until 2040 from various studies (in EJ)...... 7
Figure 1.3: Outline of the dissertation ...... 15
Figure 2.1: Monthly prices for steam coal in USD/t (CIF Eurozone, FOB Richards Bay, and FOB Newcastle) and crude oil in USD/bbl (crude oil index) between April 1996 and April 2016...... 23
Figure 2.2: Major exporters, importers, and trade flows of steam coal in 2013 and 2014...... 26
Figure 2.3: Model players in the steam coal value added chain...... 29
Figure 2.4: COALMOD-World model structure...... 30
Figure 2.5: Production cost mechanism for a model producer node...... 32
Figure 2.6: Countries included in the COALMOD-World database...... 33
Figure 2.8: Marginal cost curves (2010) for selected production nodes (in USD/GJ)...... 35
Figure 2.9: Capacity and investment costs for all production nodes in the base year...... 36
Figure 2.10: Reserves of major countries in COALMOD-World (in Gt)...... 38
Figure 2.11: Capacity and investment costs for all export nodes in the base year...... 40
Figure 2.12: FOB costs (2010) for the export countries in COALMOD-World (in USD/t)...... 42
Figure 2.13: Linear regression of average freight rates between 2002 and 2009 (in USD/t)...... 43
Figure 2.14: CIF costs in 2010 for selected routes (in USD/t)...... 44
Figure 2.15: COALMOD-World results: development of yearly global coal demand in both scenarios until 2040 (in Mtpa)...... 48
Figure 2.16: Global COALMOD-World results: aggregated consumption and imports in the Stagnation scenario (in Mtpa)...... 49
Figure 2.17: Global COALMOD-World results: aggregated consumption and imports in the 2°C scenario (in Mtpa)...... 49
Figure 2.18: Global results 2010: seaborne trade flows (in Mtpa)...... 51
Figure 2.19: Global results 2020: seaborne trade flows in both scenarios (in Mtpa)...... 52
Figure 2.20: Global results 2030: seaborne trade flows in both scenarios (in Mtpa)...... 52
Figure 2.21: Global results 2040: seaborne trade flows in both scenarios (in Mtpa)...... 53
Figure 2.22: Average prices for selected regions for all model years (in USD/t) in the Stagnation scenario...... 56
Figure 2.23: Production costs at production level for selected producers over time in the Stagnation Scenario...... 57
XIII List of Figures
Figure 2.24: Production costs at production level for selected producers over time in the Moderate Growth scenario...... 57
Figure 3.1: Illustration of different definitions of fossil fuel subsidies as a nested doll...... 66
Figure 3.2: Total steam coal consumption for different scenarios (in Mtpa)...... 75
Figure 4.1: AUS – Export tax: Australian production, consumption and exports in Base Case and in Tax AUS (in Mt)...... 86
Figure 4.2: AUS – Export tax: Changes in supply to the international market (left figure) and to domestic markets (right figure) relative to the Base Case (in Mt)...... 87
Figure 4.3: AUS – Export tax: Changes in patterns of global consumption relative to the Base Case (in Mt)...... 87
Figure 4.4: AUS – Export tax: Decomposed impact of the Australian export tax relative to the Base Case, in in Mt, and change in weighted CIF prices in percentage (right axis)...... 88
Figure 4.5: AUS – Export tax: NPV of tax revenues, in bn USD, as well as change in global CO2 emissions from coal use, in Gt (right axis), as a function of the initial export tax rate, in
USD/tCO2...... 89
Figure 4.6: Coalition – Tax revenues of the coalition’s members, in bn USD, and change in
cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of the initial
common production tax rate, in USD/tCO2...... 90
Figure 4.7: Coalition – Export tax: Decomposed impact of an export tax jointly set by the coalition of major exporters relative to the Base Case, in Mt, and change in weighted CIF prices in percentages (right axis)...... 91
Figure 4.8: Tax AUS and Tax Coalition: Comparison of average revenues per ton of CO2 abated, in
USD/tCO2, and reduction in cumulative global CO2 emissions from coal use, in Gt (right axis),
as a function of the initial value of the export tax, in USD/tCO2...... 92
Figure 4.9: AUS – Production Tax: Australian tax revenues, in bn USD, and change in cumulative
global CO2 emissions from coal use, in Gt (right axis), as a function of the initial production tax
rate, in USD/tCO2...... 93
Figure 4.10: Coalition – Production Tax: Tax revenues of the coalition’s members, in bn USD, and
change in cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of
the initial common production tax rate, in USD/tCO2...... 94
Figure 5.1: Decision tree in the CO2 disposal chain of the CCTS-Mod ...... 102
Figure 5.2: CO2 emission sources and storage potential ...... 107
Figure 5.3: Storage by sectors in MtCO2, Ref75 ...... 111
Figure 5.4: Infrastructure investment and variable costs in €bn, Ref75 ...... 112
Figure 5.5: CCTS infrastructure in 2050, Ref75 ...... 112
XIV List of Figures
Figure 5.6: CCTS infrastructure in 2050, Off75 ...... 113
Figure 5.7: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, Off75 ...... 114
Figure 5.8: CCTS infrastructure in 2050, On50 ...... 115
Figure 5.9: CCTS infrastructure in 2050, On100 ...... 116
Figure 6.1: Decision tree of the model CCTS-Mod with the option of CO2-EOR...... 122
Figure 6.2: Distribution of potential CO2 storage sites (left) and CO2 source (right) by type and volume in the data set...... 124
Figure 6.3: Captured CO2 emissions by source and storage type over time in the EU_80% scenario...... 128
Figure 6.4: Sensitivity of captured amounts over the model horizon (left side), and total costs and length of the pipeline network in 2050 (right side)...... 130
Figure 6.5: CO2 flows in the NorthSea_40% scenario in 2050 after CO2-EOR-fields are exploited. .. 135
Figure 6.6: CO2 flows in the NorthSea_80% scenario in the year 2050 after CO2-EOR fields are exploited...... 136
Figure 6.7: Cost distribution over the whole timespan in the NorthSea_80% scenario in €bn...... 136
Figure 6.8: CO2 flows in the DNNU_80% scenario in 2025 using the CO2-EOR-option (left) and in
2050 after CO2-EOR-fields are exploited (right)...... 138
Figure 7.1: Simplified network ...... 155
Figure 7.2: Electricity generation (top) and power plant investment (bottom) from 2015-2050...... 157
Figure 7.3: CO2 capture by electricity and industrial sector (area) and CO2 storage (bars) in 2015, 2030 and 2050 ...... 157
Figure A.1: Illustration of MPEC solution strategy...... 198
Figure A.2: Global trade flows in the Base Case in 2030 (in Mt)...... 200
Figure A.3: Global trade flows in AUS – Export tax in 2030 (in Mt)...... 200
Figure A.4: Global trade flows in Coalition – Export tax in 2030 (in Mt)...... 200
Figure A.5: AUS – Export tax: Optimal initial tax rates, in USD/tCO2 and the NPV of tax revenues, in bn USD (right axis), as a function of the growth rate of the export tax, in percentage per annum...... 201
Figure A.6: Coalition + USA – Export tax: NPV of tax revenues of coalition members, in bn. USD; and change in global consumption, export for all coalition members and total exports, in
percentages (right axis), as a function of the initial value of the export tax, in USD/tCO2...... 202
XV List of Figures
Figure A.7: Total supply from imports and domestic production in scenario with high estimate of reserves in operating mines (in Mtpa)...... 209
Figure A.8: Total supply from imports and domestic production in M&E scenario (in Mtpa)...... 210
Figure B.9: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, On50 ...... 215
Figure B.10: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, On100 ...... 215
Figure B.11: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, Off100 ...... 215
XVI List of Abbreviations
List of Abbreviations Abbreviation Description CCS Carbon Capture and Storage CCTS Carbon Capture, Transport, and Storage CCU Carbon Capture and Usage CfD Contracts for Differences CGE Computable General Equilibrium Model CIF Cost, insurance, freight CM Capacity Market CO Carbon monoxide CO2 Carbon dioxide COALMOD Dynamic partial equilibrium model of the world steam coal market CPF Carbon Price Floor CPS Carbon Price Support CtL Coal-to-Liquids DERA German Energy and Resource Agency (Deutsche Energie- und Rohstoffagentur) DIW Berlin German Institute for Economic Research (German: Deutsches Institut für Wirtschaftsforschung) DMO Domestic Market Obligation DOGF Depleted oil and gas fields DSM Demand-side management EC European Commission EEPR European Energy Program for Recovery EEX European energy exchange EGR Enhanced gas recovery EIA Energy Information Administration EMF Stanford Energy Modeling Forum EMR Electricity Market Reform ENTSO-E European Network of Transmission System Operators for Electricity EOR Enhanced oil recovery EPEC Equilibrium Problems with Equilibrium Constraints EPS Emissions performance standard EU European Union EU-ETS European Emissions Trading Scheme FOB Free on broad, costs include all cost incurred from the point of production to loading the coal on a ship ready for shipment. G20 Group of Twenty; is an international forum for the governments and central bank governors from 20 major economies GAMS General Algebraic Modeling System GHG Greenhouse gas GSI Geological Survey of India Gt Gigaton GW Gigawatt GWh Gigawatt hour h Hour IAEE International Association of Energy Economics IAM Integrated assessment model IEA International Energy Agency IGCC Integrated gasification combined cycle INDC Intended Nationally Determined Contribution IPCC Intergovernmental Panel on Climate Change KKT Karush-Kuhn-Tucker kW Kilowatt kWh Kilowatt hour LP Linear Problem MCP Mixed complementarity program MIP Mixed Integer Problem mn Million MPEC Mathematical Program/Problem with Equilibrium Constraints
XVII List of Abbreviations
Abbreviation Description Mt Million t MW Megawatt MWh Megawatt hour NBSC National Bureau of Statistics China NEP Grid Development Plan (German: Netzentwicklungsplan) NER300 New Entrance Reserve 300 NGO Non-governmental organization NIMBY Not in my backyard NPV Net present value NSWDPI New South Wales Department of Primary Industries O&M Operation and management O2 Oxygen OCGT Open cycle gas turbine OECD Organization for Economic Co-operation and Development OPEX Operating expenditure PESD Program on Energy and Sustainable Development ppm parts per million PRB Powder River Basin (USA) PV Photovoltaics QLDDME Queensland Department of Mines and Energy ROI Return on investment ROW Right of way SACRM South African Coal Roadmap SDG Sustainable development goals SOE State-owned enterprise SSP Shared Socio-economic Pathways t Ton TPED Total Primary Energy Demand tn Trillion TYNDP Ten Year Network Development Plan UBA Federal Environment Agency (German: Umweltbundesamt) UK United Kingdom UN United Nations UNFCCC United Nations Framework Convention on Climate Change USA Unites States of America VDKI German Association of Coal Importers (Verein der Kohlenimporteure) WEO World Energy Outlook
XVIII Chapter 1: Introduction
Chapter 1 INTRODUCTION
1.1 Motivation My research in this dissertation focuses on two closely connected topics: (1) International steam coal markets; and (2) Carbon Capture, Transport, and Storage (CCTS, or CCS). In a nutshell, it can be described as an attempt to understand the “new economics of coal” and to lift the “fig leaf” of CCTS that has been used to justify new investments in coal-fired power generation in times of full awareness of climate change.
But let me start from where I began. In 2009, when I started working on coal and its role in mitigating climate change, I did not even know that I was working on it. When I joined a research project on CCTS, my preoccupation with this technology and coal was, perhaps, typical for - at that time – a technocratic person. I was immediately fascinated by the idea of large-scale CO2-highways across Europe. At that time, I did not know much about climate change but this technological solution was appealing to me. I enjoyed the academic process of acquiring knowledge from multiple sources and independently challenging information. Working in a team with two curious and excited students – who have each perused inspiring, yet diverging, careers – we stated to learn about CCTS. Working hard to understand the economic and technological characteristics of the technology, it became clear that we would need a techno-economical model to concentrate our knowledge. While giving birth to CCTS- Mod, I learned how powerful models can be as a tool for communication but also that modeling exercises should generate “insight not numbers.” I also learned about the importance and power of language in guiding debates. While the technology was referred to as CCS in both the public and academic debate, we coined the acronym CCTS in order to emphasize the importance of CO2 transport, thus highlighting the mismatch between CO2 emission sites and the potential storage sites as a critical issue underlying the technology.
Our academic work and the trust put in us by our supervisor gave us the opportunity to discuss our insights with other researchers at an international conference. I was encouraged and impressed to see that my work was taken seriously. At the same time, I was also relieved to see that senior researchers, like the conference organizers, can sometimes also make mistakes. Having chosen the slightly misleading title of “CO2-Highways for Europe,” organizers placed us in a session where the other speakers where more concerned about cars than about power plant emissions.
Digging deeper into the literature and understanding CCTS based on my analysis, my early enthusiasm was gradually replaced with a critical view of the technology and its economic viability. After seven years of research - condensed in Part B of this dissertation - I am convinced that CCTS will not be available as a climate change mitigation technology that allows for burning coal to generate electricity and heat. 1 Chapter 1: Introduction
When studying CCTS, I became more and more interested in the nexus of climate change mitigation and possible future pathways of coal use. I realized that an understanding of coal, its characteristics, and its use would be a natural starting point to understand this - before any consideration of CCTS. While for my dissertation I observe this order and start with a thorough introduction of steam coal markets in Part A, I did not follow this logic in my academic carrier. It was only in my transition phase before officially joining the DIW Graduate Center that I started working on a project on international steam coal markets. At first I was not very keen to work on a seemingly outdated topic. Fortunately, there was again an economic model (COALMOD-World) involved that helped me to understand the market but also gave me puzzles by just not wanting to reproduce observed patterns of trade. It was my new supervisor whose expertise helped me resolve them. I understood how crucially models hinge on the underlying assumptions and how fast energy markets can change. The coal market world of 2006 - which the model was calibrated to – did not match the world of 2013, with exporters becoming importers and vice versa, and prices soaring to unprecedented heights.
Since then, the energy world continues to move. In fact, the situation of coal has again drastically changed. Our description of a resilient and low risk coal market in our 2013 work, was obsolete by the time it was published in 2015 (Holz et al. 2015). From a 2016 market perspective, international coal trade can be characterized as pressured not just by environmental policy but also cheap renewable energy sources. It also suffers from oversupply, low prices, and divestment. I prefer the perspective of international climate policy, which has finally succeeded in achieving global consensus on limiting the global temperature rise to 1.5-2°C in the Paris Agreement. It is clear that more stringent climate policies are needed to achieve the levels of emission reductions required to meet this goal. Therefore, I focus on the effectiveness of alternative climate policies and their market implications in the remainder of Part A of my dissertation. In exploring alternative approaches, I concentrate on supply- side policies for international stream coal markets.
One of the disappointing conclusions of from this research is that achieving strong reductions in coal consumption inevitably hinges on broad participation in a large number of countries in any of the examined policies. But indeed, the pressure on coal that has gained momentum and the polycentric approach with actions taken by an increasing number of countries is an encouraging sign that a global transformation of the energy system can be achieved.
The remainder of this introduction is structured as follows: the next section sheds some light on the reasons for a fundamental change in the economics of coal. Starting with a critical review of the CCTS technology, it provides an overview of institutional projections of future coal demand and gives a summary of climate policy instruments to curb fossil fuel consumption that are currently discussed in academia and public debates. Section 1.3 examines various modeling approached employed to assess the effects of climate policies and to understand underlying channels and mechanisms. Section 1.4 introduces the outline of this dissertation and gives references to chapter origins. Concluding remarks and avenues for future research are provided in Section 1.5.
2 Chapter 1: Introduction
1.2 Rationalizing the new economics of steam coal The change in the economics of coal is a slow process that is still not entirely finished. Not long ago, coal was thought to be an abundant, safe, and cheaply available energy source for industry and electricity generation worldwide. In the early years of international climate policy, e.g. when the Kyoto Protocol was signed in 1997 (UN 1997), nobody seriously challenged the role of coal at the heart of energy systems around the world. Even when the German Energiewende was set in stone in the “Energy Concept 2050” (BMWi and BMU 2010), coal was one of the pillars of the future energy system. The seeming dichotomy between rapid reduction of greenhouse gas emissions, especially from power generation and industrial processes, on the one hand, and having coal - the most emission-intensive fossil fuel - as a crucial element of the energy mix, seemed to have been resolved by a technology called Carbon Capture, Transport and Storage (CCTS).1
1.2.1 Critical view on the Carbon, Capture, Transport and Storage technology The idea that CCTS provides a solution toward a sustainable development of global energy systems emerged in the late 1990s. The general public learned about it through the Intergovernmental Panel on Climate Change (IPCC) special report on “Carbon Dioxide Capture, and Storage” (IPCC 2005). As, a one size fits all solution, the technology promised to resolve the key challenges of decarbonization by providing i) low CO2 electricity without the negative side effect of volatility of renewable energy sources; ii) a decarbonization option for energy-intensive industries; and iii) the possibility to generate so-called “negative” emissions when combining the CCTS with biomass, and generating heat and power, while sequestrating CO2 in the ground, bound by the biomass from the atmosphere (also see Gale et al. (2015a) for further details on CCTS and its development ten years after the IPCC special report).
1.2.1.1 Technical details CCTS can be spilt into three distinct technology steps. The first step comprises the Carbon Capture process. Figure 1.1 illustrate the three main technology options that are available to realize this step: Post-Combustion, Pre-combustion, and Oxyfuel (Abanades et al. 2015; Fischedick, Görner, and
Thomeczek 2015). With Post-Combustion, the CO2 is extracted from the flue gas using various solvents. It is often argued that the technology can be added to existing operations, although it entails significant thermodynamic recalibration of the system (Idem et al. 2015; Liang et al. 2015). Pre- Combustion technology relies on a pre-treatment, namely, a gasification of the fuel before combustion.
In an air and steam environment the fuel is converted into syngas, a mix of hydrogen (H2), carbon monoxide (CO), and carbon dioxide (CO2). Carbon monoxide can be further processed into additional
H2 or methane (CH4). The CO2 is then separated, leaving a hydrogen-rich fuel for further combustion
(Jansen et al. 2015). For the third alternative, fuel is burned in a pure oxygen (O2) and CO2 environment. With no impurities in the combustion atmosphere (especially no nitrogen), the
1 While the technology was initially labeled Carbon Capture and Storage (CCS), the label CCTS emphasizes that CO2 transport is a critical component, as detailed in the following section.
3 Chapter 1: Introduction
combustion products are CO2 and water vapor (H2O), which can then be separated (Stanger et al. 2015). Pre-treatment of fuel, air separation or thermodynamic reconfiguration, all entail substantial reductions in process efficiency, typically up to one-third of reference levels. Moreover, they come with additional upfront- and operational costs (Fischedick, Görner, and Thomeczek 2015; Rubin, Davison, and Herzog 2015).
Figure 1.1: The process chain of carbon capture, transport, and storage.
Source: Adapted from Oei (2016), based on Fischedick et al. (2015).
In the next step of the CCTS technology, captured CO2 needs to be transported from emission source to a final storage destination, as on-site storage is not usually available. CO2 conditioned to a super- critical state can be transported like natural gas or crude oil. While tankers are an option for transporting smaller quantities on direct links, pipeline transport is more economic for bulk transport (Geske, Berghout, and van den Broek 2015a; Geske, Berghout, and van den Broek 2015b). Once investment costs for pipeline infrastructure are borne, which account for the major share of transport costs, operating compression stations and maintenance make up the low variable costs (Fischedick, Görner, and Thomeczek 2015; Oei, Herold, and Mendelevitch 2014).
A suitable reservoir for long-term CO2 storage is a permeable structures that can accommodate the
CO2 sealed by non-permeable rock (Herold, Oei, and Tissen 2011; Krevor et al. 2015). Candidate storage sites include oil fields, gas fields, saline aquifers, and coal beds (Bachu 2015; Birkholzer,
Oldenburg, and Zhou 2015). While CO2 injection has been practiced in the oil and gas industries since the mid-1990s, experience with respect to the long-term environmental impact and permanent storage are limited (Jenkins, Chadwick, and Hovorka 2015; D. G. Jones et al. 2015). Due to high case specificity, estimates of storage potential and associated costs come have high uncertainty. Given public opposition to on-shore storage in many European countries, the remaining option is offshore
4 Chapter 1: Introduction
storage, which is more expensive and less researched (see Hirschhausen, Herold, and Oei (2012), Ashworth et al., (2015) and Chapter 6 for more details).
1.2.1.2 High hopes and realities While critical voices against CCTS were present in the early days of climate change mitigation discourse, they were overshadowed by those reluctant to turn their backs on fossil fuels, at least in the mid-term, viewing CCTS a bridging technology (Martínez Arranz 2015). Against early claims of its unsustainability, high costs, and long lead times, the technology had mainstreamed by 2010 and was available from the EC (e.g. the European Energy Program for Recovery (EEPR) (EC 2009) and the New Entrance Reserve (NER300) (EC 2010)) but also from other governments.2 Despite available funding and high hopes put into CCTS, the technology has not matured (Reiner 2016). Large-scale implementation of the CCTS value chain is still not proven (2015b), while the projects setup for demonstration from 2008 to 2011 have been canceled or delayed indefinitely (MIT 2016; Oei et al. 2014a). Various reasons for the failure of CCTS, or what Hirschhausen, Herold, and Oei (2012c) call “a lost decade” for CCTS, have been identified (see e.g. Hirschhausen, Herold, and Oei 2012c; Scott et al. 2013; Marshall 2016). Low CO2 prices have created little incentive for incumbents in the energy industry to invest in the technology.3 Moreover, a wrong technology and industry choice is theorized with a focus on electricity instead of industrial processes, and equal funding for differently mature CO2 capture technologies. In addition, the costs and complexity of CO2 transport and storage have been neglected, thus emphasizing that Carbon Capture, Transport and Storage (CCTS) would be the more appropriate abbreviation. Perhaps most importantly, the competition from mature, low-cost renewables has been underestimated. Hirschhausen et al. (2013) identify overly optimistic assumptions regarding technological learning and resulting cost reductions for CCTS, while, at the same time, respective projections for renewables have not been corrected downwards. The authors find that the latter to be the main driver for high shares of CCTS in various institutional projections. The model presented and applied in the second part of my dissertation can be seen as a building block to understand why the technology has not lived up to its expectations based on costs and readily available alternatives.
Recognizing the lack of progress in CCTS technology development, it has been removed from the regular IEA “Projected Costs of Generating Electricity” publication (IEA, NEA, and OECD 2015). Still, CCTS is not entirely off the table. It is again the IPCC and the Integrated Assessment Model (IAM) community who shows that mitigation costs would increase by 25% to 300% without CCTS and, in total, only 4 out of 11 models were able to achieve targeted emission reductions for a 2°C scenario without this technology option (Edenhofer et al. 2014). In its 450 ppm scenario, the IEA World Energy
Outlook (IEA 2015a) estimates that 4 Gt of CO2 will be captured using CCTS in 2040 (60% from the electricity sector and 40% from industry). Those projects that have moved forward against the general
2 Examples are found in, e.g. Marshall (2016) with details on funding available in Australia. A list of U.S. funding for CCTS by the U.S. DoE is available at U.S. DoE (2016): “Carbon Capture and Storage Research | Department of Energy.” Energy.gov Office of Fossil Energy. http://energy.gov/fe/science-innovation/carbon-capture-and- storage-research. 3 Such a situation was also observed with other pollution control innovations in the past (cf. Hackett 1995).
5 Chapter 1: Introduction
trend in CCTS are all associated with enhance oil recovery using CO2 injection (CO2-EOR)
(Mendelevitch 2014). This cannot be considered a CO2 mitigation technology, as it serves producing an additional source of emissions and substantial amount of CO2 resurfaces which the oil (Gale et al. 2015b). China, and India, as the world’s largest coal consumers, would have a natural interest in developing CCTS, instead China is pursuing very modest research, while support is lacking in India (Viebahn, Vallentin, and Höller 2014; GCCSI 2015b).
1.2.2 A changing environment for coal While CCTS has not seen progress, international climate policy succeeded in achieving global consensus on the urgent need to combat anthropogenic climate change at Conference of the Parties (COP)21 in Paris (UN 2015b). Consequently, pressure on coal increased. The estimates of fossil fuels that must remain in the ground in order to achieve the 2°C target and to prevent irreversible atmospheric changes put the heaviest burden on coal (Meinshausen et al. 2009). Between 82% and 88% of current coal reserves are considered “unburnable,” compared to 33%-35% of oil and 49%-52% of gas reserves (McGlade and Ekins 2015). It is not only scientists (e.g., Johnson et al. 2015) and NGOs (e.g., D. Jones and Gutmann 2015) that deduct the need to phase-out coal if climate change mitigation is taken seriously, but also by policy makers. Actions are being taken globally: The Obama Administration enacted the Clean Power Plan (EIA 2015b), and China introduce a moratorium on new coal power plants and mines (The State Council of the People’s Republic of China 2016). Seven smaller EU countries are already coal-free: Belgium, Cyprus, Estonia, Latvia, Lithuania, Luxemburg, and Malta. Portugal is planning to phase-out coal in 2020, followed by Finland in the 2020s. The UK announced plans to phase-out coal before 2025; actions also planned by Austria and Denmark (Hill 2016; Jacobsen 2014).
Despite claims that the world cannot do without coal (Umbach 2015), the industry is starting to feel the fading perspective for coal. Caught between the shale gas boom and decreasing costs of renewables, by 2016 numerous U.S. coal producers (including Peabody Energy Cooperation, Arch Coal Inc., and Alpha Natural Resources, listed first, second and fourth out of the top five (measured by output) U.S. coal mining companies) have filed for bankruptcy (Sussams and Grant 2015, 18; Mooney and Mufson 2016; EIA 2016b), while in 2013, a total of 271 mines were idled or closed (EIA 2015c). Steam coal production declined by around 20% between 2005 and 2013 (though preliminary data for 2014 shows a very slight increase in production, of around 1% compared to 2013) (IEA 2015c, IV.424). A large share of producers in Queensland, one of Australia’s two main coal regions, produced coal at loss in 2014 (McCracken 2015). Coal companies world-wide are not only affected by low prices, but are also challenged by the divestment movement, with big financial institutions recognizing the danger of climate change. Financial institutions acknowledge the entailing carbon investment bubble, making current fossil assets likely to become stranded with more stringent climate policies in place (Leaton et al. 2013). Therefore, they reject further financing of fossil fuel projects (Arabella Advisors 2015). Observing this massive failure of classical business models, Fulton et al. (2015) suggest an alternative paradigm of evaluating and managing company risk, taking into account non-linear and structural changes in both in fossil fuel markets and finance.
6 Chapter 1: Introduction
1.2.3 Wedge of global institutional projections for coal demand Against the background of the decline of the coal industry, the spread of the projections of coal demand4 from a range of institutions, energy companies, and scientific papers, as shown in Figure 1.2, may be somewhat puzzling, at first glance. The spread illustrates the uncertainty around future coal consumption that still prevails. It is important to note that, although the forecasts are provided by a seemingly heterogeneous group. some common motives can still be identified. BP, ExxonMobil, and Statoil are all oil and gas companies with an inherent incentive to prolong the unconstrained use of fossil fuels and will be negatively affected if stringent climate policies are enforced. International Energy Agency (IEA) and U.S. Energy Information Administration (EIA) are both reported to constantly lag behind with their assessment of the potential of renewable energy sources and vice versa, favoring projections that do not foresee any structural changes to the energy systems (Metayer, Breyer, and Fell 2015; Gilbert and Sovacool 2016).
Figure 1.2: Projected coal demand until 2040 from various studies (in EJ).5 Source: Own illustration based on BP (2016), EIA (2016a), ExxonMobil (2016), IEA (2015a), McGlade and Ekins (2015), MIT (2015), and Statoil (2016).
4 Including steam coal, metallurgical coal, and lignite. 5 The underlying models provide estimates in five to ten year steps. Therefore, the line between these steps are only for illustrative purposes. IEA WEO (2015) - CPS refers to the IEA World Energy Outlook’s Current Policy Scenario (IEA 2015a), NPS stands for New Policies Scenario, and 450ppm is the 450 ppm scenario; Statoil - REFS refers to the Statoil World Energy Perspectives Reform Scenario (Statoil 2016), RENS to the Renewables Scenario, and RIVS to the Rivalry Scenario; EIA EIO – RC refers to the EIA Energy International Outlook’s Reference Case (EIA 2016a), and M&E refers to the extraction path for coal calculated to be consistent with a 2°C target by McGlade and Ekins (2015), excluding the option of CCTS.
7 Chapter 1: Introduction
While some of the institutions provide several scenarios describing different future, they also need to be treated with care. For example, in the case of IEA WEO 450ppm and M&E projections, both claim to be consistent with a 2°C target. Still there is a divergence of 60 EJ, more than a third of current consumption, also for this two scenarios. The main reason for the disagreement is the crucial difference in the role that CCTS plays in the respective future energy system. While the IEA WEO 450pmm scenario assumes that 75% of installed coal-fired power generation capacity is equipped with CCTS, M&E estimate coal consumption patterns that would result in the absence of this technology. Given substantial doubts whether the technology will ever enhance from or even achieve the demonstration phase (cf. 0 of this dissertation), the M&E scenario is the only one that projects a coal demand pattern which is robustly in line with the 2°C target.
Taking a look at the projections label as reference or baseline scenarios, they are all project future coal demand in the range of 150EJ to 190EJ, by 2040. Even if these scenarios underestimate the pace of deployment of renewables and give favor to fossil fuel; a gap of 66% and more to be consistent with a 2°C target clearly means that current climate change policy regimes are not stringent enough to establish confidence in a decline in future coal consumption. Therefore, existing regimes need to be strengthen and alternative polices need to be introduced.
1.2.4 Instruments of climate policy Generally, climate policy instruments can be organized along different metrics.6 One common metric is to sort them according to the targeted side of the market for emission-intensive goods (in the scope of this thesis steam coal): those policies targeting the consumers are referred to as demand-side policies, while those addressing production are referred to as supply-side policies. Another frequent metric is to order the policies according to their general approach: Market-based economic instruments can take the form of taxes/subsidies, or tradable allowances/credits, as well as border tax adjustments and are directed to influence the decision-making of profit-maximizing firms (Kolstad et al. 2014, 364). While these instruments try to incentivize a desired behavior of private actors, additional approaches can be implemented if the conduct is not satisficing or there is a lack of political feasibility for a particular economic instrument. Purely regulatory approaches are alternative economic instruments. They establish a rule (standard) that a regulated entity has to comply with, while threatened with a penalty in case of non-compliance. Hybrids of regulatory and market-based approaches are widely applied (incentive regulation). Besides, information programs and governmental provision of goods and services can be used as additional instruments to achieve climate objectives. Table 1.1 provides an overview of different types of climate policies, differentiated along the two metrics discussed above. The Grantham Research Institute maintains a database of global climate legislation which details different policies that have been implemented (Grantham Research Institute 2015a).
Each policy has its specific advantages and disadvantages. Typical policy evaluation criteria assess the efficiency, the effectiveness, and the feasibility of a policy intervention (Perman et al. 2012). These
6 This section heavily builds on earlier work published as Collins, and Mendelevitch (2015).
8 Chapter 1: Introduction
criteria can be used in assessing and improving policies ex-ante, or they can be used to verify results, withdraw inefficient policies and correct policy performance ex-post (Kolstad et al. 2014, 238).
When evaluating policy instruments, differences in the economic structure, institutions, and policy objectives between high-income and low-income countries need to be accounted for (Kolstad et al. 2014, 242). Low-income countries might lack the right market structures to underpin cost-effectiveness (e.g., liberalized energy markets). Furthermore, they are likely to incur higher opportunity cost of capital and of governmental resources. Consequently, they face higher institutional and other transaction costs which reduces feasibility for certain policy designs, and calls for capacity building and financial support.
Table 1.1: Examples for different types of supply-side and demand-side climate policies. Supply-side policy Demand-side policy Economic Resource production tax Carbon or fuel use taxes instruments – Resource export taxes Border carbon price adjustments taxes Taxes on fossil fuel capital (income) Economic Removal of fossil fuel producer subsidies Removal of fossil fuel consumer instruments – subsidies subsidies Renewable energy subsidies Economic Cap-and-trade for production rights Cap-and-trade for consumption rights instruments – Offsets for leaving assets in ground Emission reduction credits or offsets tradable allowances and credits Regulatory Prohibiting development of certain Coal plant emission standards approaches resources or use of certain technologies Building codes Limiting production or export (e.g. a via quota) Comprehensive emissions assessment in environmental impact review of new fossil fuel supply projects Government Restricted leasing of state-owned Infrastructure expansion (district provision of lands/waters for coal, oil and gas heating / cooling; electric vehicle goods and development. charging station; wind transmission) services Decision to not develop specific resources Policies to restrict export credit agency or infrastructure (oil pipelines and or multilateral development finance for terminals; coal ports, etc.) coal power stations Funding to compensate resource owners for leaving reserves undeveloped Policies to restrict export credit agency or multilateral development finance for coal mining and other supply infrastructure Information Divestment by institutions/companies Energy audits programs, involved in fossil fuel production Vehicle or appliance labelling voluntary Extraction-based emissions accounting by Territorial emissions accounting actions, and nations and sub-national governments; other life-cycle based accounting of embedded GHGs in fossil fuels sold in marketplace Source: Lazarus, Erickson, and Tempest (2015).
1.2.4.1 Demand-side policies
Demand-side policies to reduce CO2 emissions, which provide indirect incentives to reduce fossil fuel consumption, have received most attention in the academic literature and are widely used in practice.
9 Chapter 1: Introduction
For instance, carbon pricing instruments place an explicit price on emissions – either directly, as a carbon tax, or indirectly, through a cap-and-trade scheme (OECD 2013a). Such instruments have been implemented (or are scheduled to be implemented) in 39 countries, and at the jurisdictional level in a further three countries (Kossoy et al. 2015a, 22).
In theory, carbon taxation performs well in terms of economic efficiency. Assuming the correct tax level, it can be broadly applied to cover all sectors, technologies and fuels, and can be used to trigger changes in production, consumption behavior and choices with the use of existing institutions. Intertemporal economic efficiency from an optimal tax trajectory can come with the co-benefit of creating a path dependency with less fuel use resulting in less political opposition. The associated transaction costs and administrative burden is relatively low, if the tax is levied at the point of production or entry into a country (upstream) and incorporated into the existing tax system (Somanathan et al. 2014, 1159). However, the high visibility of a tax puts a burden on the political feasibility. The general public is less likely to accept such instruments and can be easily (mis)guided by influential fossil fuel lobbies. The distributional incidence of a carbon tax depends on the revenue recycling mechanism. For the case of vehicle fuel taxes, Sterner (2012) shows that their effect is neutral in Europe and can be weakly regressive (before revenue recycling) in other rich countries, but it is generally progressive in least developed and developing countries such as India, Indonesia, China, and many African countries. To the contrary, in the case of a subsidy removal, the fear of social unrest may restrain political feasibility.
There are many other policy instruments which generate an implicit carbon price through regulatory intervention. Prominent examples are emissions performance standards, minimum flexibility requirements, renewable portfolio obligations (see Oei et al. (2014b) for a discussion of regulatory options to reduce CO2 emission in the power sector, and Chapter 7 for some numerical applications to the UK). Other demand-side policies include measures that promote energy efficiency and reduced energy consumption (as discussed in articles in Economics of Energy & Environmental Policy Symposium on “Energy Efficiency”: Gandhi et al. 2016; R. Hahn and Metcalfe 2016; Rosenow et al. 2016; Houde and Spurlock 2016). Efficiency regulation is often criticized to come with a pronounced direct and indirect rebound effect (increase in energy-consuming activity due to lower marginal energy requirement (direct rebound effect), and economic-growth induced increase in energy consumption due to a decrease in the cost of energy services (indirect rebound effect), first observed by Jevons (1865). Regulatory approaches are very diverse in nature and no general findings can be derived to assess these measured against environmental policy evaluation criteria. They are used across many countries, and may serve other goals in addition to reducing emissions (OECD 2013a).
Under the right conditions, market-based carbon pricing instruments are theoretically the most efficient policy instruments to reduce emissions (Stavins 2003). Practical outcomes appear to support theory, with cap-and-trade schemes and carbon taxes found to drive more abatement at given costs compared to other policy instruments (OECD 2013b). However, the effectiveness of these instruments may be undermined by inadequate design and implementation. For instance, prior to the commencement of the European Emissions Trading Scheme (EU-ETS) it was predicted that its
10 Chapter 1: Introduction
effectiveness would be reduced due to suboptimal permit allocations (Kemfert, Diekmann, and Ziesing 2004), and this has now eventuated in practice (Ellerman, Valero, and Zaklan 2015; Neuhoff et al. 2015). Consequently, the EU-ETS and similar economic instruments in place worldwide have only generated low carbon prices (averaging 5 EUR/tCO2 in 2014 (IEA 2015b, 23)). In contrast, higher carbon prices are needed to initiate a process of substituting away from coal in the power generation sector. For instance, one recent estimate of the price that would drive coal-to-gas switching in Europe is around 40 EUR/tCO2 (Gray 2015, 49). Moreover, it is likely that even higher carbon prices are necessary to drive the closure of old, fully-depreciated, coal-fired generators (IEA 2014a, 17; Hecking 2016).
In the absence of full participation in a global climate policy, demand-side policies are prone to cause carbon leakage: emissions-intensive activities shift to non-participating countries, such that emission reductions in the participating countries are partly offset by emissions increases in the non- participating countries (see e.g. Felder and Rutherford 1993; Sinn 2008). Richter (2015) provides an overview of empirical studies of the carbon leakage effect, which is undisputed in its existence, but controversial in its magnitude. Ex-ante, the supply elasticity of coal is found to be crucial for the magnitude of the effect, with higher elasticity leading to stronger leakage effects (Burniaux and Oliveira Martins 2012). Using General Equilibrium frameworks that incorporate the interaction between trade and the environment, most studies find only moderate rates of leakage (Felder and Rutherford 1993; Paltsev 2001; Di Maria and Werf 2008). High rates of carbon leakage are estimated by Babiker (2005), who criticizes overly simplistic assumptions on market and industry structure. Employing an integrated assessment framework, Arroyo-Currás et al. (2015) identify limited leakage of 15%, if the U.S., the EU and China act as pioneer regions. In an ex-post empirical study of the effect of the Kyoto Protocol on GHG emissions, Aichele and Felbermayr (2015) find a change in the production patterns of emission-intensive goods and thereby evidence for carbon leakage of about 8%. Combined with earlier findings by Aichele and Felbermayr (2012), the carbon leakage rate is estimated at roughly 40%.
To close the gap for leakage and to arrive at a global carbon policy in international climate change negotiation, authors of the Economics of Energy & Environmental Policy Symposium on "International Climate Negotiations" (Gollier and Tirole 2015; Weitzman 2015; Cramton, Ockenfels, and Stoft 2015) argue in favor of a uniform carbon price. By contrast, Stiglitz (2015) proposes a more flexible and differentiated approach including carbon taxes, a system of cap-and-trade, and regulatory mechanisms but also a green fund to account for shared responsibility.
A “green paradox” has also been theorized, where the expectation of future demand-side policies induces resource producers to increase their present rates of extraction in order to maximize net present value (Sinn 2015). For coal, Haftendorn, Kemfert, and Holz (2012) suggest that in practice the green paradox may not be relevant, while Bauer et al. (Bauer et al. 2013) find a short term reduction of coal prices due to stringent climate policy. Gerlagh (2011) argues that the green paradox relies on oversimplified model assumptions with total depletion of the resource and high substitutability between
11 Chapter 1: Introduction
different energy fuels. Hoel (2012) adds that the paradox is only prevailing if policy targets low cost suppliers while it is absent if it affects mainly high-cost suppliers of fossil fuel.
1.2.4.2 Supply-side policies Supply-side policies represent an alternative and more direct route to address negative effects of fossil fuel combustion. One important factor to consider when making a choice whether to introduce a demand-side or a supply-side policy is the ratio of the elasticity of demand to that of supply, as this ratio drives the leakage risk for either policy. Lazarus and Ericson and Tempest (2015) calculate this ratio for different fuels and regions based on various studies. They find mixed evidence for supply-side and demand-side leakage risk for coal. Collier and Venables (2014) make the case that, in the absence of full participation in a global climate policy, a targeted supply-side policy will be more effective in reducing emissions from coal combustion than a demand-side policy. In particular, carbon leakage is minimized under a supply-side policy rather than a demand-side policy if the price elasticity of demand is high relative to the price elasticity of supply – which is considered to be the case for coal in the long-run (Collier and Venables 2014). The threat of a green paradox is also thought to be eliminated with a properly designed supply-side policy – in particular, one that targets high-cost coal deposits for closure (Hoel 2013). Another benefit of supply-side policies is that they achieve predictable and observable outcomes with low transaction costs (Collier and Venables 2014). It has also been suggested that supply-side climate policies may drive greater emission reductions for a given marginal cost, and will limit over-supply of fossil fuels and associated “carbon lock-in” effects (Lazarus, Erickson, and Tempest 2015).
An important consideration in the context of supply-side policies is whether producers should be compensated for a loss of profits associated with the fossil fuel that is then not produced anymore. A number of studies suggest that under a policy of freely allocated depletion quotas, enhanced scarcity rents for fossil fuels that are extracted can offset the loss in profits (Eisenack, Edenhofer, and Kalkuhl 2012; Kalkuhl and Brecha 2013; Asheim 2013). Similar results are obtained for a policy which confiscates fossil fuel reserves (Asheim 2013). These findings indicate that there is no need for overall compensation for foregone profits. However, there may still be a need for compensation payments between producers to alleviate internal distributional effects, whereby producers with low-extraction- cost reserves will benefit at the expense of other producers (Asheim 2013).
One type of supply-side policy acts to directly remove coal reserves from production – whether to a partial extent (focusing on high-extraction-cost reserves for economic efficiency) (Harstad 2012), or more extensive, the progressive closure of the entire coal industry (Collier and Venables 2014). Another type of supply-side policy is a depletion tax (or alternatively, a depletion quota), which is analogous to the demand-side policy of a carbon tax (or for a depletion quota, a carbon budget). For instance, in Chapter 4 we propose a tax on the energy content of steam coal, levied by a coalition of major coal exporters. Modelling shows that a tax levied by a coalition of major coal exporters is preferred to a tax levied by a single major coal exporter, and that a production tax generates better outcomes than an export tax (though they note a production tax is likely to be politically contentious). A supply-side policy for coal could also take the form of an export-licensing regime adopted by a
12 Chapter 1: Introduction
coalition of major coal exporters, in analogy to the existing safeguards regime for uranium exports; based on the reasoning that the regulation of commodity exports on the basis of their harmful or unethical end-use is a widely accepted principle, and should be extended to coal (A. Martin 2014). Lazarus, Erickson, and Tempest (2015) provide a comprehensive taxonomy of supply-side climate policies.
To date, there has been limited experience with the implementation of supply-side policies. The concept of preserving fossil fuel reserves has some precedent in the Yasuni-ITT Initiative, which was a proposal by the Ecuadorian government in 2007 to preserve oil reserves, but ultimately was not carried through (P. L. Martin 2014). A recent initiative that directly targets future coal supply is the “No New Coal Mines” campaign. It was started by the President of Kiribati who urged the leaders of the world to support this call for a moratorium on new and expansion of existing mines (Tong 2015). The policy is supported inter alia by the Obama administration (Warrick and Eilperin 2016) and by the Australia Institute and argues in favor of a global moratorium on new coal mines (Denniss 2015b). Another supply-side policy which is broadly discussed at least since 1997 (cf. World Bank 1997) only fragmentally implemented is a removal of fossil fuel subsidies. Chapter 3 discusses the introduction of the latter two policies and finds no significant effect on consumption for the former. For the latter a reduction in consumption of coal that is consistent with a 2°C target could be achieved depending on the specification of the policy.
1.3 Mathematic modeling of policy incentives and games on international resource markets
While the instruments listed in section 0 all serve the ultimate goal of reducing anthropogenic CO2 emissions, they all use quite different channels to do so. To be able to capture the characteristics of a particular channel and even more so to be able to quantify interdependencies with technological and regulatory constraints, one needs to rely on the appropriate modeling framework. The most comprehensive approach is taken by IAMs. Seminal work is done by Nordhaus (1991), and Edmonds and Reilly (1985). These models try to combine knowledge from natural science on the effect of CO2 emissions on climate change and the environment, with a comprehensive representation of economic activity that is the origin of anthropogenic CO2 emissions, and which is situated in this very environment (Nordhaus 2013). In advanced models of this class, there is also a feedback between current CO2 emissions and future economic activity. While these models are widely applied to assess the effects of climate policies, they are also criticized for providing overly confident results used to guide policy debates while relying on uncertain assumptions (Pindyck 2015). Just like Computable General Equilibrium Models (CGE) these models employ a top-down approach where individual economic actors are aggregated to industries and representative agents. Therefore, these models are not able to capture characteristics of market power and different risk attitudes. They also inherently lack a detailed representation of technologies (e.g., intermittency of renewable sources), though current model versions try to overcome this shortcoming (Johnson et al. 2016). Moreover, substitutability between different fuels and technologies is modeled in a static and reduced form, limiting their ability to assess the transformability of energy systems.
13 Chapter 1: Introduction
By contrast, energy system models are able to deliberately consider how these systems can change over time and how climate polices influence the speed and scope of transformations (Pfenninger, Hawkes, and Keirstead 2014). These technology-oriented models focus on the energy conversion system, on the demand-side (e.g., efficiency measures) as well as supply side (e.g., wide range of generation technologies). Prominent model examples are PRIMES (Capros et al. 1998), TIMES- MARKAL (Fishbone and Abilock 1981), and EFOM (Finon 1979). While these models convey the “big picture” from the energy system perspective, they again lack a representation of strategic interactions on involved energy markets.
Partial equilibrium models can capture the economics of particular fuel markets by accounting for both, technical infrastructure restrictions, and strategic behavior by individual actors. Therefore, they are able to assess incentives induced by climate policy on the level of different strategic actors. A large stream of literature applies this model setup to different fossil fuel market, for example oil (e.g., Huppmann and Holz 2012), and natural gas markets (e.g., Egging, Holz, and Gabriel 2010), and electricity markets (e.g., Hobbs and Rijkers 2004). COALMOD-World, the model used in Chapter 2 and Chapter 3, also belongs to the same class of models, all set up as mixed complementarity problems (MCP). Recently, Huppmann and Egging (2014) combine the rich presentation of individual fuel markets into a comprehensive framework to allow for endogenous fuel substitution. A similar approach is taken in Chapter 7, where a detailed model of the electricity sector is combined with a comprehensive representation of the CCTS technology chain also applied in emission-intensive industries.
MCPs can only account for two basic non-cooperative market settings: perfect competition and oligopolies with competition à la Cournot (Cournot 1838) resulting in a Nash-Cournot equilibrium. But often strategic interaction is more complicated. For example, it can involve hierarchical constellations with one dominant player anticipating reactions of other market participants, for example in a Stackelberg Leader-Follower (Stackelberg 1934) setting. Computationally, this can be represented by a Mathematical Program under Equilibrium Constraints (MPEC). This setup can also be used to represent situations where climate policies are set on a country level, anticipating the reaction of the respective industry (see Chapter 4 for an application and Siddiqui (2011) for solution concepts). For the case of binary options, modeling techniques are currently an active field of research (see, e.g., Huppmann and Siddiqui 2015). Other strategic setups involve multi-stage hierarchies with oligopolies on each of the stages represented by Equilibrium Problems under Equilibrium Constraints (EPEC) (Huppmann and Egerer 2015).
While these approaches cover a wide range of market, technological, and regulatory settings, they still cannot account for the particularities of specific technologies and policy instruments. CCTS, just like natural gas and oil, constitutes a network industry, where economies of scale are an inherent feature that cannot be represented by the previously described techniques due to underlying non-convexity. Integer programming can be used to integrate this characteristic, as it is done in Chapter 5.
In general, all modeling results can only be interpreted in the context of the applied framework. While models are powerful tools to aggregate knowledge and data in a structured and explicit way,
14 Chapter 1: Introduction
interpreting their results requires just as much understanding of the mechanics of the examined markets and interactions. As emphasized by leading scholars: “models yield insights, not numbers” (cf. e.g., Hamming 1962; Huntington, Weyant, and Sweeney 1982).
1.4 Outline of the dissertation The outline of the dissertation is divided into two parts (see Figure 1.3): Part A consists of Chapters 2-4, which all deal with the international market for steam coal. While Chapter 2 introduces the model, using and extending it, the two subsequent chapters analyze the effects of different supply- side climate policies on the market. Part B is dedicated to CCTS. Chapter 5 presents the modeling framework and scenarios of European-wide deployment. Chapter 6 adds CO2-EOR as an important component to be considered when evaluating the technology and examines scenarios of regional CCTS implementation. Chapter 7 combines the insights from electricity market modeling with a detailed representation of CCTS in the power and industrial sector. As a showcase, it tests the effect of different policy options on the resulting electricity mix in the UK.
Figure 1.3: Outline of the dissertation
1.4.1 Part A: Modeling the effects of climate policies in the world steam coal market The effect of climate policies on the world steam coal market is modeled in Chapters 2-4. Chapter 2 introduces a model of the international steam coal market. The model is applied to study production subsidy removal and a mine moratorium policy in Chapter 3. In Chapter 4, the initial framework is extended and applied to analyze the effect of export and production taxes.
Chapter 2 presents a tool that is able to replicates global patterns of coal supply, demand and international trade in great detail. It features endogenous investments in production and transportation
15 Chapter 1: Introduction
capacities in a multi-period framework and represents the substitution relation between imports and domestic production of steam coal. It simulates production, trade, price, and capacity development and can readily be applied to evaluate policy implications through scenario analysis. Two strongly diverging scenarios are examined to showcase the functionalities of the modeling framework, but also to discuss its limitations.
I examine two further supply-side policies in Chapter 3. The first is a production subsidy reform introduced in major coal producing countries, in line with the G20 initiative to reduce global fossil fuel subsidies. The second policy tests the effects of a globally implemented moratorium on new coal mines, as suggested by many scholars. Both scenarios require substantial data collection on subsidy levels and coal reserves in operating mines, which I both present in this chapter. Scenario results show that a subsidy removal, while associated with a small positive total welfare effect, only leads to an insignificant reduction of global emissions. By contrast, a mine moratorium induces a much more pronounced reduction in global coal consumption due to limited availability and a strong price increase. Depending on the specification of reserves, the moratorium can achieve a coal supply path consistent with the 1.5-2°C target.
The initial modeling framework is extended to a two-level setting in Chapter 4. Finding solutions to large-scale two-level problems with an optimization problem at the upper, and an equilibrium problem at the lower level is a current field of research in Operations Research. This chapter applies state-of- the-art solution methods to examine the effects of introducing export and, alternatively, production taxes by different groups of players on the steam coal market. While the policy comes with a double- dividend of rent creation and emission reduction, we find that a unilateral export tax on steam coal has little impact on global emissions and coal prices, as other countries compensate for reduced export volumes from the taxing country. By contrast, a tax jointly levied by a coalition of major coal exporters would significantly reduce global emissions from steam coal, and raise significant revenue. Production taxes consistently yield higher tax revenues and have greater effects on coal consumption and emissions associated with smaller rates of carbon leakages.
1.4.2 Part B: The vision of CCTS as low-carbon solution for the electricity and industry sector
The potential contribution of CCTS to the decarbonization of the electricity and industry sector is calculated in a numerical, model-based analysis in Chapter 5, followed by a more sophisticated model including the option of CO2 enhanced oil recovery (CO2-EOR) in Chapter 6. Chapter 7 aims at closing the research gap between electricity market models, which do not put any emphasis on CCTS, and models of CCTS infrastructure development, which neglect how the technology is driven by decisions in the electricity market.
Chapter 5 presents a mixed-integer, multi-period, welfare-optimizing network model for Europe, called
CCTS-Mod, used to analyze the economics of CCTS in the wake of expected rising CO2 prices. The model incorporates endogenous decisions on carbon capture, pipeline and storage investments, as well as capture, flow and injection quantities based on given costs, CO2 prices, storage capacities, and point source emissions. Given full information about future costs of CCTS technology and CO2 prices, 16 Chapter 1: Introduction
the model determines a cost minimizing strategy on whether to purchase CO2 certificates, or to abate the CO2 through investments into a CCTS-chain on a site by site basis. We apply the model to analyze different scenarios for the deployment of CCTS in Europe. For example, under high and low CO2 prices, respectively. CCTS can contribute to the decarbonization of Europe’s industry sectors (in particular iron and steel, and cement industry), as long as the assumption of sufficient on- or offshore storage capacities holds. The power sector has higher capture costs and invests in the CCTS technology at higher CO2 certificate prices than the industry.
An improved data set of costs in Chapter 6 reveals more realistic insights, as early cost projections turned out to be too low. The chapter analyzes the layout and costs of a potential CO2 infrastructure in Europe over the time horizon up to 2050 based on a critical review of the current state of the CCTS technology adding the important option of combining CCTS with CO2 enhanced oil recovery. The technology cannot be considered emissions reducing due to the oil and significant amounts of CO2 that are extracted. Still it can change the economics of CCTS projects by generating an additional revenue stream for oil production. The potential of CO2-EOR is found to be limited and concentrated in the North Sea in Europe. Regardless of the CO2-EOR option, the iron and steel sector is found to have the lowest capture costs, starting CCTS deployment once the CO2 certificate prices exceed 50 €/tCO2.
The cement sector starts investing at a threshold of 75 €/tCO2, followed by the electricity sector when prices exceed 100 €/tCO2. The degree of CCTS deployment is found to be more sensitive to variable costs of CO2 capture than to investment costs. Moreover, scattered CCTS deployment increases the unit cost of transport and storage infrastructure by 30% or more.
Finally, Chapter 7 presents a two-sector electricity-CO2 (ELCO) modeling framework, which closes the gap between electricity sector and CCTS-focused models. Players can invest in various types of generation technologies including renewables, nuclear, and CCTS. The detailed representation of
CCTS also comprises also industry players (iron and steel as well as cement), as well as CO2 transport and storage including the option for CO2-EOR. The model also simulates interactions of the energy-only market with different forms of national policy measures. All players maximize their expected profits based on variable, fixed, and investment costs as well as the price of electricity, CO2 abatement costs, and other incentives, subject to technical and environmental constraints. Demand is inelastic and represented via a selection of type hours. The model framework allows for regional disaggregation and features simplified electricity and CO2 pipeline networks. The model is balanced via a market clearing for the electricity as well as the CO2 market. The equilibrium solution is subject to constraints on CO2 emissions and renewable generation share. The model is applied to a case study of the UK Electricity Market Reform to illustrate the mechanisms and potential results attained from the model.
17 Chapter 1: Introduction
1.4.3 Chapter origins and own contribution Table 1.2 displays the pre-publications and further information on the individual contribution for each chapter of the dissertation.
Table 1.2: Chapter origins
Dissertation Chapters Pre-Publications Own contribution 2. The COALMOD-World The Global Coal Market - Supplying Joint work with Franziska Holz, Model: Coal Markets until the Major Fuel for Emerging Clemens Haftendorn and Christian 2040 Economies. Cambridge, UK: von Hirschhausen. Roman Cambridge University Press. Mendelevitch had the lead role in the model development, the Forthcoming as DIW Data
implementation in GAMS, and the Documentation 85. Berlin: DIW Berlin. analysis of results. 3. Testing Supply-Side Submitted to Climatic Change Single author paper Climate Policies for the Forthcoming as DIW Berlin Global Steam Coal Market – Can They Curb Discussion Paper No. 1604, 09/2016 Coal Consumption? 4. Coal Taxes as Supply- Submitted to Journal of Joint work with Philipp M. Richter Side Climate Policy: A Environmental and Resource and Frank Jotzo. Roman PartCoal A: Markets Rationale for Major Economics Mendelevitch and Philipp M. Richter Exporters? jointly had the lead role in the model DIW Berlin Discussion Paper No. development, the implementation in 1471, 04/2015, Berlin. GAMS, and the analysis of results. Philipp M. Richter had the lead role in the writing of the manuscript. 5. Modeling a Carbon Environmental Modeling and Joint work with Johannes Herold Capture, Transport, and Assessment 05/2014; December and Pao-Yu Oei. Roman Storage Infrastructure for 2014, Vol. 19, Issue 6, pp 515-531 Mendelevitch and Pao-Yu Oei jointly Europe developed the model, and its DIW Berlin Discussion Paper No. implementation in GAMS. Andreas 1052, 09/2010, Berlin. Tissen was also involved in developing a first draft of the model. The writing of the manuscript was executed jointly. 6. European Scenarios of The Energy Journal, Vol. 37, SI3, pp Joint work with Pao-Yu Oei. Roman CO2 Infrastructure 171-194 Mendelevitch and Pao-Yu Oei jointly Investment until 2050 – developed the model and its Resource Markets Working Paper CO2-Enhanced Oil implementation in GAMS. Pao-Yu WP-RM-36 at University of Potsdam, Oei had the lead in analyzing the Recovery Keeps the 2013. Mirage Alive political setting for CCTS in the EU. Roman Mendelevitch had the lead in collecting data on CO2-EOR, and analyzing the results. The writing of the manuscript was executed jointly. 7. The Impact of Policy Under review in Energy Systems; Joint work with Pao-Yu Oei. Roman Measures on Future Mendelevitch and Pao-Yu Oei jointly DIW Berlin Discussion Paper No. Power Generation developed the model and its Portfolio and 1521, 11/2015, Berlin. implementation in GAMS. Roman
Part Carbon, B: Storage and Transport, Capture, Infrastructure – A IEEE Conference Publication for the Mendelevitch had the lead in Combined Electricity and 12th International Conference on the collecting data. Pao-Yu Oei was in CCTS Investment and European Energy Market (EEM), May charge of the implementation of the Dispatch Model 2015. UK case study. The writing of the manuscript was executed jointly.
18 Chapter 1: Introduction
1.5 Concluding remarks and outlook for future research The COP21 Paris agreement has brought about a clear commitment to reduce anthropogenic greenhouse gas (GHG) emissions to a level that will most likely keep the increase of global mean temperature below 2°C and striving for 1.5°C. Achieving this goal entails a fast transition from our currently fossil fuel based energy systems towards renewable energy sources. As a consequence, this renders large shares of current fossil fuel reserves unburnable (Meinshausen et al. 2009). For coal, as the most emission-intensive and most abundant fossil fuel, this means that over 80% of current reserves will need to stay in the ground (McGlade and Ekins 2015). Carbon, Capture, Transport, and Storage has once been thought of as allowing combining climate mitigation while prolonging the use of coal. The technology has not lived up to the high hopes put in it and did not deliver proof of concept at a demonstration scale (Reiner 2016). Findings from Chapter 5 and Chapter 6 of this dissertation suggest that CCTS can be an option for energy-intensive industries where CO2 can be captured at lower costs compared to power generation. It will not play a major role in decarbonizing the electricity sector, even if it is coupled with CO2-enhanced oil recovery to generate additional revenue.
One of the main reasons for this is the growing availability of cheap renewable sourced alternatives for electricity generation. Already in 2013, levelized costs of electricity for onshore wind and solar PV, caught up with those of new coal-fired installations in Europe (see e.g., Schröder et al. 2013). Some institutions also list nuclear energy as a valid alternative to fossil fueled electricity generation (see e.g., IPCC 2014; EC 2013b). Even a shortage of uranium has been hypothesized due to a future large scale deployment. While the shortage of uranium is just a question of prices and not of physical quantities (Mendelevitch and Dang 2016), due to high investment costs and unresolved issues of final storage, the technology cannot be considered a promising alternative technology (Lévêque 2014; Schneider et al. 2016). Even emerging economies with cheap domestic fossil fuel and uranium resources like Kazakhstan, are trying to develop their renewable potentials and embark on a green growth path (Egerer, Mendelevitch, and Hirschhausen 2014).
At the same time, fossil fuel markets, and at the forefront the market for coal, are currently on an uncertain track. “New economics of coal” are characterized by low prices, oversupply, and divestment. Regulatory approaches that ultimately push coal out of the electricity mix are applied in an increasing number of high-income economies which traditionally relied on coal. Nevertheless, current policy efforts - that until now mostly concentrate on the demand-side of fossil fuels - are insufficient to achieve required CO2 emission reductions. Supply-side climate policy presents an alternative approach that directly targets fossil fuel production and reserves. Due to induced scarcity rents, these often come with a double dividend of reducing emissions while increasing prices and thereby compensation for foregone revenues. Using a large-scale partial equilibrium model of the world steam coal market developed in Chapter 2of this dissertation, I examine the effects of three different supply- side climate polices in Chapter 3 and Chapter 4. While there is currently a lot of talking about fossil fuel subsidies in the context of sustainable development goals and the G20, I find that the removal of production subsidy for coal has negligible impact on coal consumption. Importantly, this effect critical hinges on the subsidy definition and on estimated current subsidy levels. An interesting extension of my analysis would be to include underpriced externalities into the definition of subsidies. This is 19 Chapter 1: Introduction
current practice by IMF and yield subsidy estimates that are several orders of magnitude larger than those of OECD. Two alternative supply-side policies might be to introduce export, or production taxes on coal, or to adapt a moratorium on new coal mines. Just like with demand-side approaches, to induce desired reductions in coal consumption, these policies require a broad implementation. The stability of these coalitions and the distributional effects induces by these policies are promising avenues for further research at the nexus of economic and mathematical research.
In contrast to coal, natural gas and oil are also heavily used outside the electricity sector, namely in residential heating and transportation. Achieving the 2°C goal also requires transforming these sectors towards renewable energy sources. One approach that is propagated as a solution to this problem is the coupling of different sectors, that is, coupling electricity and heating, and electricity and transportation. Battery-electric vehicles can provide an alternative technology for motorized private transport, but also for public urban transport (Kunith, Mendelevitch, and Göhlich 2016). Heat pumps are one option of coupling the electricity and the heating sector. The electrification of these sectors, however, increases the demand for carbon-free electricity and therefore increases the pressure on phasing-out coal.
Therefore, more effective climate policies are needed to break the inertia inherent in current energy systems and phase-out fossil fuels at the required pace. Based on past experience with global climate policy, however, achieving a first-best solution with an overarching coalition seems unrealistic (Cole 2015). Ostrom (2010) suggests a polycentric approach with additional policy schemes for various regions and energy sectors. With the momentum gained on reducing coal consumption a next important step is to reconsider the role of the oil and gas sector in transnational environmental agreements. On the surface, the logical strategy would be to hinder stringent climate protection regimes altogether, as observed with the oil and gas industry over the course of the COP negotiations (van den Hove, Le Menestrel, and de Bettignies 2002; Downie 2015). Once arguing away climate change is not an option anymore, alternative strategies of fossil fuel industries can be: i) collusion or claims for compensation and economic diversification or ii) investment in efficiency to reduce costs and increased competition (Van de Graaf and Verbruggen 2015).
However, in my opinion, there might also be another alternative: if climate policy primarily targets coal, then a window of opportunity opens up for oil, and especially for gas, to substitute for coal in electricity generation. Therefore, there might be a level of stringency of climate regimes that is actually favorable for the oil and gas industry. Exploring the level of climate actions that would be in line with the interests of non-solid fossil fuels is a promising avenue for further research. This question can be examined using multi-fuel partial equilibrium models (e.g., Huppmann and Egging 2014; Hagem and Storrøsten 2016) or as well as with the help of analytical frameworks that account for incentives of different transnational actors (e.g., Hagen, Kähler, and Eisenack 2016). The knowledge gained from this research can be used to align incentives and create a consensus on climate protection actions, but also to identify potential hidden agendas.
But also for the future of coal and CCTS economic research will keep playing a vital role in analyzing balanced phase-out paths and critically assessing the role of technologies in achieving climate targets.
20
Part A COAL MARKETS
21 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Chapter 2 THE COALMOD-WORLD MODEL: COAL MARKETS UNTIL 2040
2.1 Introduction: A comprehensive coal market model is needed Developing a structural model of international steam coal markets is a true challenge in an energy world at the crossroads, where the last years have been characterized by some fundamental structural changes and the future perspectives are unclear.7 On the one hand, there is a strong trend toward renewables in the OECD (Organization for Economic Cooperation and Development) countries, in particular in Europe but also in China, where climate change, and local pollution awareness has induced a number of climate policies. In parallel, the availability of cheap shale gas and low-cost wind and solar power have accelerated the decline of the coal industry in the US. On the other hand, South- East Asian countries have seen increasing demand for coal fueling their economic development. In a situation where the future role of coal in the global energy mix is put in question, projections of global coal demand exhibit a large spread, indicating fundamental uncertainties on its future development.
Irrespective of the global demand trend, international coal trade has an important role to play in current and future global coal markets. Although the share of internationally traded coal in total world coal consumption is relatively small – approximately 15% (IEA 2013) – it has a major impact on the evolution of domestic markets and therefore plays a central role in the analysis. Some scholars argue by many scholars, international market prices are set by the arbitrage between Chinese coal delivered from the main production regions in the North to the consumption centers in the South-East versus coal imported to these demand centers (see e.g., Morse and He 2015).
When coal market modeling came back to the surface of the academic literature, after its first “boom” in the 1980s, coal markets had seen some stable years with continuous low prices, more or less competitive trading relations and a stable increase in seaborne trade (Haftendorn and Holz 2010). However, coal markets have been caught in the turmoil of a general energy price and cost increase since 2007. In 2008, they experienced a price peak similar to the oil markets (see Figure 2.1), and they compete with other mining sectors for qualified labor, mining services, and machinery, which has led to a substantial cost increase in many producing countries. The strong demand increase driven by Asia and the subsequent capacity expansions have led to a consistent cost escalation affecting the
7 This chapter is forthcoming as DIW Data Documentation 85, 11/2016 (Holz et al. 2016). A previous version was published as Chapter 10 in the Book: The Global Coal Market - Supplying the Major Fuel for Emerging Economies, by Morse and Thurber (Holz et al. 2015).
22 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
fundamentals of the coal markets. However, the last three years were characterized by plummeting prices, again. The two main driving factors behind this trend are overcapacity originating from overly optimistic demand projections, and a stark decline in oil price which makes up a significant part of production and transportation costs of coal (see Figure 2.1).
With coal at the center of discussions between climate change mitigation policies and economic development, a model of the international steam coal market can deliver insights into the mechanics of the market and assess the implications of climate policy measures in a comprehensive manner. Besides, other issues like security of supply or market power abuse might gain importance again, underlining the need for a practical assessment tool. This report presents the functionalities of the COALMOD-World model (cf. Haftendorn, Holz, and Hirschhausen 2012; Holz et al. 2015) which replicates global patterns of coal supply, demand and international trade in great detail. It features endogenous investments in production and transportation capacities in a multi-period framework and represents the substitution relation between imports and domestic production of steam coal. It simulates production, trade, price, and capacity development and can readily be applied to discover policy implications through scenario analysis.
Figure 2.1: Monthly prices for steam coal in USD/t (CIF Eurozone, FOB Richards Bay, and FOB Newcastle) and crude oil in USD/bbl (crude oil index) between April 1996 and April 2016. Source: HWWI commodity prices in the Thompson Reuters Datastream database. The remainder of this report is organized as follows: the next section provides an introduction to the international steam coal market and gives an indication of the uncertainty about future demand. The evolution of the COALMOD modeling framework and how it is embedded in the literature on steam coal market models is given in Section 2.3. Section 0 provides a detailed description of the model
23 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
structure. The explicit mathematical formulation and details on model sets, parameters, variables and equations, as well as variations of the COALMOD framework to investigate various policy and market issues are collected in Appendix A. Section 0 presents input data including information on data sources. Using two scenarios, Section 2.6 illustrates the functionalities for the model. Section 2.7 discusses limitations of the model framework, and Section 2.8 concludes.
2.2 The international steam coal market
2.2.1 Types of coal Coal is commonly categorized as steam coal, metallurgical coal or lignite, based on its material properties and end-use. Steam coal is the set of coal types that are typically combusted to produce steam8. In 2014, Around 70% of steam coal was used to produce electricity and heat, and the remainder mostly for other industrial heat-consuming activities (IEA 2015c, III.68). IEA (2015c, I.25) defines steam coal as anthracite, other bituminous and sub-bituminous coal, with an energy content ranging from 20 GJ/t to as much as 30 GJ/t (IEA 2015c, I.25).
Steam coal is mined at either surface or underground mines, mainly depending on the depth of the coal seam (Speight 2012). The raw coal is processed through crushing, screening, and beneficiation/washing operations to meet customer specifications. To transport the coal to ports or markets, rail is most common, but river barges are also used (as well as other modes of transport over short distances). Where necessary along the supply chain, coal is stored in open air stockpiles or enclosed silos.
2.2.2 Coal markets Large-scale demand for steam coal originated in the eighteenth and nineteenth centuries, where its use in powering steam engines was central to the industrial revolution and subsequent economic growth in Europe and the United States (Fernihough and O’Rourke 2014; Chandler 1972). By the beginning of the 20th century coal had become the dominant source of energy worldwide, though during the early to mid-20th century it lost shares to oil and gas (Smil 2000). The oil crises of the 1970s triggered the revival of the steam coal market, as countries which had previously imported large quantities of oil for power generation sought to bolster their energy security by diversifying their power supply (IEA 1997, 25). Coal was a substitute for oil due to its wide abundance and low cost (Thurber and Morse 2015a, 12–13). From 1980 to 2000, steam coal consumption grew steadily in most OECD and non-OECD regions alike (with Europe being an exception), and from 2000 to 2005 there was a large spike in steam coal consumption in non-OECD countries, in particular in China and the rest of the Asia-Pacific region (IEA 2014b).
Figure 2.2 provides an overview of major steam coal producers and consumers in 2014. Since 2005, steam coal consumption in the OECD has decreased by around 12% (IEA 2015c), due to general
8 Metallurgical coal is bituminous coal which is used to produce coke for use in the iron and steel industry. Lignite is a low-quality brown coal which is also used to produce steam.
24 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
trends of decarbonization and lower energy consumption (IEA 2014c, 172). However, over that same period consumption has continued to grow in non-OECD countries – by 10 times the volume of the OECD decrease (IEA 2015c). This rapid growth in demand triggered significant investment in supply capacity and transport infrastructure (IEA 2014c, 186). However, in the past few years demand growth has slowed.
Since the 1980s, China has been the world’s largest consumer of steam coal. India was the world’s third largest steam coal consumer since 1995, but since 2005 has almost doubled its consumption to become the world’s second largest steam coal consumer in 2014 (on a tonnage basis) – narrowly overtaking the USA, whose consumption has decreased by around 20% over the past decade (IEA 2015c, III.30-III.32). Other large consumers of steam coal over the past two decades are South Africa, Japan and the Russian Federation; while in the 1970s and 1980s, Poland, the United Kingdom, and Germany were also in the mix. In more recent years, analysis by Steckel et al. (2015a) shows that it is not only China and India which are driving a renaissance of coal; rather, it is gaining dominance in numerous developing countries, especially in South-East Asia but also in Turkey.
Table 2.1: Major steam coal producers and consumers in 2014.. Major producers in 2014 Major consumers in 2014 China (3,200 Mt) China (3,280 Mt) United States (770 Mt) India (760 Mt) India (560 Mt) United States (750 Mt) Indonesia (470 Mt) South Africa (174 Mt) South Africa (250 Mt) Japan (137 Mt) Australia (246) Korea (100 Mt) Russian Federation (190 Mt) Russian Federation (77 Mt) Kazakhstan (94 Mt) Kazakhstan (67 Mt) Colombia (84 Mt) Poland (60 Mt) Poland (61 Mt) Indonesia (60 Mt) World production 6,150 Mt World consumption 6,090 Mt Source: IEA (2015b). The world’s largest consumers of steam coal are also its largest producers. Since the mid-1980s (when it overtook the USA), China has produced the largest volumes of steam coal, followed by the USA. India has been the world’s third-largest producer of steam coal since the 1990s, having overtaken South Africa (IEA 2015c, III.10-III.11). Along with Australia and the Russian Federation, these countries account for over 90% of world steam coal production – with China alone accounting for 52% of the total. Similar to consumption trends, Poland, the United Kingdom and Germany were historically large producers of steam coal, but by the 1990s had lost any significant market share.
Figure 2.2 depicts major importers, exporters and trade flows of steam coal in 2013 and 2014. Worldwide, the total quantity of internationally traded steam coal in 2014 represented 17% of total demand, with the majority being seaborne trade (IEA 2015b, III.39, III.44, III.49). The total volume traded has increased at an average annual rate of 6% between 1990 and 2014, and the proportion of seaborne trade increased at an average annual rate of 2% over the same period. For most of the 1990s and 2000s, Japan and Korea were the world’s largest importers of steam coal. However, since the late 2000s, China, and subsequently India, overtook Japan as the world’s largest importers. Indonesia, Australia, and the Russian Federation are the world’s largest exporters of steam coal,
25 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
followed by Columbia and South Africa. Due to their geographical location, South Africa, as well as Russia, are “swing suppliers”, which export to both the Pacific and Atlantic regions according to market dynamics (IEA/OECD 2014, 50).
Figure 2.2: Major exporters, importers, and trade flows of steam coal in 2013 and 2014. Source: OECD/IEA (2015).
2.3 Literature
2.3.1 State of the international literature An extensive review of the – sparse – coal-market specific modeling literature until 2010 is provided in Haftendorn and Holz (2010) and Paulus and Trüby (2011). There were some early modeling efforts applied to the US and international coal markets in the 1970s and 1980s (see Shapiro and White 1982; Kolstad and Abbey 1984). Often, coal modules are part of larger energy system models, as in most of the models applied in the Energy Modeling Forum number 2, “Coal in transition 1980 – 2000” (EMF 1978). However, both the situation on the international steam coal market as well as modeling techniques have evolved since the 1980s. For other energy and resource markets, such as natural gas, multi-period models with endogenous investments have been developed during the 2000s (e.g., Hartley and Medlock 2006; Egging, Holz, and Gabriel 2010, for world natural gas markets).
Two modeling teams have been the major contributors to the recent renaissance of coal market modeling applying the equilibrium technique, coincidentally both from Germany where steam coal imports have traditionally had an important role: a team from Energiewirtschaftliches Institut an der Universität zu Köln (ewi), and the developers of the COALMOD framework. For the former, seminal papers are by Trüby and Paulus (e.g., 2012) who developed a one-period trade model to test the market structure until 2008. They also developed a multi-period model which they use to investigate the interaction between the Chinese and the world coal market (Paulus and Trüby 2011). A coal market model focusing only on Chinese coal supply was developed by Rioux et al. (2015).
26 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
2.3.2 Development of the COALMOD model framework and publications The COALMOD model framework was developed and continuously extended by a joint team of researchers at the Department of Energy, Transport and Environment at DIW Berlin, in close cooperation with researchers from Berlin University of Technology (Workgroup for Infrastructure Policy, WIP). The first version was developed by Haftendorn and Holz (Haftendorn and Holz 2010). This version called COALMOD-Trade focused on steam coal exports, only, and calculated static market equilibria for 2005 and 2006, to access market structure and compare model specifications. The model was also introduced to a broader, non-academic audience in two articles, (in German, Haftendorn et al. 2011; Haftendorn et al. 2012)). Further model development was undertaken within the framework of the “Global Coal Market” project led by the Program on Energy and Sustainable Development (PESD) at Stanford University. PESD invited the modeling team Berlin to expand their model of the seaborne export market (Haftendorn and Holz 2010) into a comprehensive model of the global steam coal market that additionally includes national and regional trade flows. Haftendorn, Holz, and Hirschhausen (2012) presents COALMOD-World, which incorporates these features and constitutes a dynamic model of the international steam coal market with a 2030 horizon, and 2006 as the base year. Moreover, it introduces the sophisticated mine-mortality mechanism and depletion of reserves to the model. Holz et al. (2015) updates the base year of the model to 2010, taking into account major shifts in cost and demand. The model presented in this report is fundamentally based on Holz et al. (2015) with minor extension of global coverage and extending the model horizon to 2050.
The ongoing development and application of the COALMOD framework has led to a considerable number of publications. Haftendorn (2012) provides an overview of publications using the COALMOD modeling framework until 2012. Similarly, Haftendorn et al. (2013) summarizes scenario results obtained using the COALMOD-World model until 2013. Currently, there are two main strains of applications of the COALMOD framework:
The first is concerned with analyzing the market structure of the steam coal market. The COALMOD-Trade model was set up to test for market power abuse in 2005 and 2006. Haftendorn and Holz (2010) in general find no evidence for market power exertion, but rather show evidence for spatial price discrimination. These findings are put into broader perspective in Haftendorn, Hirschhausen, and Holz (2008), adding the observation of increasing market concentration by a small number of dominant firms, and calling for close surveillance through regulatory authorities. The latter issue is further investigated by Haftendorn (2012). The paper explores the hypothesis that the incumbent dominant firms located in South Africa and Colombia withheld supply to the European market in 2004 and 2005, which allowed a new entrant, namely Russia in the market. Findings suggest that market power was exerted but no collusion between the incumbents was detected. The second strain is concerned with the effect of short-term and long-term policy and other shocks to the market. Hirschhausen et al. (2011) COALMOD-Trade model was used to examine the issue of security of supply in Europe. They find little risk from supply disruption and market power exertion. Haftendorn, Kemfert and Holz (2012) use COALMOD-World to examine 27 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
interactions between climate policies and the global steam coal market until 2030. They examine a unilateral European climate policy which is found to induce high leakage. By contrast, a supply- side policy like an Indonesian export-limiting is reported to be most effective in an environment of low intensity of global climate policy when the market is constrained. A third scenario investigates a fast-roll out of Carbon Capture, Transport, and Storage (CCTS). If the technology is realized, increased demand due to reduced efficiency can lead to additional positive climate effects, if the market is constrained. A different approach is taken by Richter, Mendelevitch, and Jotzo (2015), who explore the complementarity between export taxes and climate change mitigation. They find that only for large coalitions of exporters the double dividend of market power rents and
significant reduction in CO2 emissions from steam coal use can be realized, while for smaller
coalition there is high leakage. A moderate global CO2 tax can achieve the same outcome but comes with different distributional implications. Mendelevitch (2016b) focuses on the effect of supply-side climate policies on the steam coal market. For the removal of coal producer subsidies
it finds only insignificant reductions in coal-based CO2 emissions. By contrast, a moratorium on new coal mines is found to have a substantial impact on future coal consumption, coming with the side-effect of increased prices which can offset foregone profits from new mines. Mendelevitch (2016a) provides an overview of various supply-side climate policy scenarios and applications performed with the COALMOD framework.
2.4 The COALMOD-World model
2.4.1 Overview COALMOD-World is a multi-period model that simulates market outcomes, trade flows, and prices for the period 2010 to 2040 in five year steps, as well as investments in the coal sector value chain. The model assumes profit-maximizing players who optimize their expected and discounted profit over the total model horizon. The model result is a cost-efficient outcome that abstracts from the short-term real world frictions and cycles but gives a valuable indication of future trends.
The value chain of the steam coal sector will be reflected in the model setup. Various types of players are involved at each stage. Producers can be large national and sometimes state-owned companies. There are a few large multinational coal companies but also many small companies, usually operating in one country only. Transport infrastructure in the production countries can be built by the mining company or by another entity. Often, it consists of rail infrastructure but in some countries trucks or river barges are used. Export ports can be dedicated to one company or utilized by multiple companies. Traders as intermediaries also play a role in this market that is characterized by bilateral relationships; they can be vertically integrated or contractually connected to every stage of the industry.
2.4.2 Model structure COALMOD-World is a multi-period equilibrium model of the global steam coal market with two types of players: producers and exporters facing consumers represented by a demand function. The stages of the real-world value chain that are included in the model are represented in Figure 2.3 by small
28 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
rectangles inside the larger producer and exporter boxes. Coal import terminals and the subsequent land transport links to the final consumers are excluded because their capacities are assumed to be sufficient. By assumption, demand for seaborne import coal is situated close to the import port. The second type of demand node can be reached by a land link directly from the producer. The producer player includes the coal mining company and also the land transport links. The exporter operates the export terminal and also pays for the seaborne transport. All players are aggregated on a national or regional (subnational) level.
The model producers and exporters represent stylized players defined for aggregated production, export and consumption nodes primarily determined using geographical parameters. A production node represents a geographically restricted area (mining basin) and aggregates the mining companies present in that area into one player called “producer.” In the model, production node and producer are equivalent terms. Production nodes are defined based on the geography of reserves, type of coal, and production cost characteristics.
An export node represents the coal export terminal of one region and aggregates the real world coal export harbors present in that region into one model player called “exporter.” Here again, export node and exporter are used synonymously. The export nodes are primarily defined based on geographic factors. A demand node represents a geographic area where the coal is consumed. It aggregates the consumption by the coal-fired power plants in a region. It can have access to seaborne coal through a port or only be supplied domestically. The demand nodes are primarily defined based on geographic factors, but other factors may come into play such as the connection to a port or the presence of mine- mouth power plants.
Figure 2.3: Model players in the steam coal value added chain.
29 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.4 shows the model structure and the relationships between producers, exporters and demand. The producers extract and treat (i.e., produce) the coal under some production capacity constraint. They can sell it either to local demand nodes or to linked exporters. They bear the production and the inland transport costs. Further, they can invest in additional production capacities and in transport capacities to local demand or to the exporter. These investments are, in turn, also subject to constraints. Moreover, a constraint on reserves is applied over the total model horizon. Shadow prices (dual variables) are obtained for all constraints and may indicate an incentive to expanding capacities.
In the same logic as the producers, the exporters maximize their profit. Each exporter is linked to a maximum of one producer. The profit for each year is defined by the revenue from sales net of the costs of purchasing the coal at the FOB price from the producer, the costs of operating the export terminal, the costs of transport (shipping) to the final market and finally the potential costs of investing in additional export capacity. An exporter can only sell to a demand node with a port. Since each model exporter has a dedicated model producer, the energy content factor of the exporter is equal to the energy content of the producer that supplies it for any given year.
Final demand is located at a consuming node and represented by a linear inverse demand function, that is, a marginal willingness to pay function. Individual demand functions for each demand node are constructed using respective reference prices and reference demand values for the model starting year 2010 and using demand growth projections for future years. Moreover, we make assumptions about the demand elasticities. The producers can be in indirect contact with the final demand through their exporter or sell directly to their domestic demand. Prices are expressed in USD per GJ because we concentrate on the demand for energy embodied in the coal.
Figure 2.4: COALMOD-World model structure. 30 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
The model runs until 2040 and calculates yearly equilibria for the energy quantities sold in the years 2010, 2015, 2020, 2025, 2030, 2035, and 2040, which can be called “model years.’' Also, the players can decide on investments in each model year that will be available in the next model year.9 Thus, the model does not only calculate an equilibrium within each model year but also over the total model horizon regarding optimal investments. For the years between the model years (e.g., for 2011, 2012, 2013, 2014, between 2010 and 2015), we interpolate the produced quantities since they are necessary to model the reserve depletion. We assume that production and other capacities will be made gradually available in the years between the model years to reach their new value in the following model year.
Both producers’ and exporters’ problems are profit maximization problems over the entire model horizon. The players have perfect foresight, meaning that they choose the optimal quantities to be supplied in each model period and the investments between model periods under the assumption of perfect information about current and future demand. Thus, the model simulates how demand should be served optimally given that the players behave rationally using all the information that is available to them.
It is important to note that the traded quantities are the quantities of energy contained in the coal and expressed in petajoules (PJ). Whenever the model needs to deal with mass quantities in million tons of coal (for the costs, capacities and investments) these energy quantities are converted to mass using a conversion factor in tons per gigajoule (GJ) that varies by producer (see Figure 2.4).
We assume short-run production cost functions for a year for each producer that can vary over time. Until recently, resource market models have often used the same short-run costs for every model period (e.g., Egging, Holz, and Gabriel 2010). This is not a realistic solution for a model of the coal market since there are many potential factors that influence future costs and change the short run costs. Other models only use the long run marginal costs (e.g., Lise, Hobbs, and van Oostvoorn 2008). This is also problematic for a model of the coal market since the short-term marginal costs determine the prices in each period and, as we have seen in our previous static modeling work (Haftendorn and Holz 2010), enable us to represent the trade flows accurately.
Figure 2.5a shows the logic of aggregation of individual mines in a mining basin to form the model producers’ marginal cost curves. We assume that a specific mine with a certain geological setting operates at constant marginal costs. The horizontal line together with the dashed line represent the reserves of a mine. The horizontal line represents the production capacity at a given point in time. Thus, in order to obtain the aggregated cost curve in one period we add the production capacities on the q-axis and connect them with their respective marginal costs on the mc-axis.
9 Although equilibria are actually calculated until 2050, we interpret the results only until 2040 because of a risk of distortion of the investment results given the short payback period after 2040.
31 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.5: Production cost mechanism for a model producer node. After this static consideration, let us consider how this cost function might evolve over time. This effect of cumulated production is illustrated in Figure 2.5b. We assume that, even if all the mines along the cost curve may produce coal in one period, the cheapest mines are depleted first. Thus, we follow the rules stated by Hotelling (1931) that for exhaustible resources the cheapest deposits are extracted first. This is due to the fact that the cheapest mines are usually the deposits that are the easiest to access and are, hence, the oldest ones in operation. The effect of cumulated production from one model period to another is then that the cheapest mines are mined out and that the cheapest producer in Figure 2.5a disappears from the cost curve. We call this effect “mine mortality”.
Mine mortality causes the intercept of the cost function to increase as shown in Figure 2.5b. The mine mortality factor gives the position on the cost curve of the previous period to determine the new intercept. Graphically, this is the passage from Figure 2.5a to Figure 2.5b. If the mine mortality factor is equal to one (i.e., the mine mortality rate is equal to 100%) it means that the cumulated production leads to a complete depletion of the oldest, cheapest mines. This may happen for mature and old mining basins. On the contrary, a factor close to zero means that the mines situated on the low cost segment of the basin’s cost curve still have significant reserves and will only be depleted in the mid- to long term.10 Since the slope of the cost curve indicates the relation of cheap and more expensive mines in the production region, the slope is the second factor in determining the new intercept’s position.
The explicit mathematical formulation and details on model sets, parameters, variables and equations, as well as variations of the COALMOD-World framework to investigate various policy and market issues are collected in Appendix A.
10 For the coal sector, in which reserve assessments are sparse, another direction of the change of the cost function intercept is possible: the intercept could decrease due to the discovery and opening of lower-cost mines. In Haftendorn et al. (2012), we discuss this possibility and propose an extended endogenous cost mechanism. Moreover, technological progress in coal extraction methods could reduce costs over time even in existing mines. However, we lack empirical data to include these mechanisms for cost decreases in the model structure.
32 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
2.5 Model specification and input data In section 0 we introduced the concepts of nodes and model players. The model simulates the market on an aggregated basis in that we do not include individual mines or coal-fired power plants separately. However, the spatial characteristics of the market and the associated transport costs make it necessary to define aggregated nodes in the different producing and consuming countries. Section 2.5.1 describes our choice of model countries and nodes, and then we provide a detailed overview of our data in subsequent sections. The data collection required a major effort since there is no central source available. We generally collected data from publicly available sources; however, there is scarcity of data in the public domain and improvements could be achieved by using more detailed data. In order to remove inconsistencies in the data and to properly represent the base year 2010 we carried out a calibration of the data. Thereby, the COALMOD-World database is able to provide realistic runs and give insights into the future developments of the global steam coal market as is shown in Section 2.6.
2.5.1 Countries and nodes definition We include all countries that were either consuming at least 5 million tons per annum (Mtpa) or producing and exporting at least 5 Mtpa in 2010, at the time of the development of the model. Some additional countries that are becoming relevant players on the global market are included too (e.g., Mongolia and Mozambique). The world map in Figure 2.6 shows the represented countries, indicating their role on the world steam coal market (importer, exporter, or both).
Figure 2.6: Countries included in the COALMOD-World database. In our data set, we distinguish production and consumption nodes. Hence, a country that only produces for export is represented in the data set with a production node from which it also exports (e.g., Colombia). A country that only imports and consumes coal is included with a consumption node (e.g., Italy and Turkey). For a country in which production takes place and that also consumes coal, we include at least one production node and one consumption node. For larger countries, there can be more than just one production and demand node; this is the case for the US, China, India, and Australia. The complete list of countries and nodes in the model can be found in Table A.4 in the Appendix.
33 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Producing nodes are generally defined by mining basins that are restricted by geological realities. The location of power plants is more dispersed as it relates to human settlement patterns. This makes it more difficult to locate our consuming nodes. For the consumers that can only be reached via an import harbor we define the demand as being located close to the port. For consumers that can be reached by land we aggregate regional data on capacities to form the demand node and define an average for the transport costs.
2.5.2 Production, costs, and reserves The cost data is based on Baruya (2007) as well as several recent IEA sources and publications by the German Association of Coal Importers (VDKI). For each export country, Baruya (2007) provides estimates for the low and high average costs; the other sources usually follow the same methodology. This information is used to construct the producers’ cost functions for the base year. We assume that the average low and high costs also represent unit costs for the cheapest and the most expensive mine. We construct a marginal cost function using the low estimate to determine the curve intercept. We place the second point at the intersection of the high cost estimate and the maximum production capacity in order to obtain a linear marginal cost curve.
In Figure 2.7 we depict the marginal cost curves for the base year 2010. One can see that the different production regions in a large country (e.g., in China) can well have different production costs. As would be expected, Poland is the most expensive producer on the world market, while some Chinese regions (Shanxi, Shaanxi, Inner Mongolia – abbreviated SIS), the Powder River Basin (US), and South Africa are on the lower end.
Clearly, there has been a substantial cost escalation in the last years that is evident when comparing to our earlier 2006 base year (Haftendorn, Holz, and Hirschhausen 2012). However, this cost increase did not hit all producers in the same way and proportionally across regions. While some producers such as Australia and also Indonesia have experienced substantial cost increases, others like the US Powder River Basin have almost maintained their costs of four years earlier. This has led to a shift in the relative cost structure: for example, with increasing export capacities (c.f. Figure 2.7) Powder River Basin coal is competitive with all other producers in all consuming regions, while the more expensive Australian coal is only competitive to ship within its Pacific home region (see, e.g., IEA 2011; VDKI 2013).
For the producers from the CIS (Russia, Kazakhstan, and Ukraine), Colombia, Venezuela, South Africa, and Indonesia, the cost data and the parameters of the marginal cost function as well as the capacities are updated based on IEA (2011). Countries with more than one production node require more detailed data on production capacities.11 This data is used to build a merit order curve of costs
11 For the United States, EIA (2012) gives production information; for China, data comes from the NBSC (2007); for India, from Indiastat (2012, Spreadsheet “Average Cost of Production of Coal”) and Indian Bureau of Mines (20112012, in the chapter title “Prices”); and for Australia, updated data from Rademacher (2008, 78), Bayer (2012), NSWDPI (2009), and QLDDME (2009) are used. For Vietnam, the production capacity is taken from Rademacher (2008); since there was no cost data available for Vietnam, these were determined using relevant price data. For Poland, the costs are based on Ritschel and Schiffer (2007).
34 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
using production data for each model producer and so estimate the producers’ cost functions. In order to determine the cost functions in the long run, some assumptions had to be made about mine mortality in each five-year period and the associated rate of growth in the intercept. Table A.6 in the Appendix shows these assumptions and the values of the intercept and slope for each producer.
Figure 2.7: Marginal cost curves (2010) for selected production nodes (in USD/GJ). Source: Authors’ work based on EIA (2012, table 32); NSWDPI (2009); QLDDME (2009); Rademacher (2008); Baruya (2007); NBSC (2007); Ritschel and Schiffer (2007). The investment costs are a major input to the multi-period model since they determine the investment decisions. For the value-added chain from production to the export terminal, the IEA estimates investments costs of 50 USD (2007) per ton of annual capacity addition (USD/tpa) and for some new projects this number goes up as high as 80 USD/tpa (OECD and IEA 2008a). Rademacher (2008) finds average investment costs of 62 USD/tpa with a wide range from 15 USD/tpa for some Australian opencast mines to 130 USD/tpa for new underground mines in Ukraine and Mozambique. But investment costs in Australia can also exceed 100 USD/tpa if the project includes new transport and washing facilities.
We therefore assign values from 40 to 80 USD/tpa to the different producers’ investment costs for production capacity based on information about the country and mine mortality rates. The assignment is based on factors such as the prevalent type of mining, geology, and the state of technology. Unit investment costs and the production capacity for the base year and every production node are shown in Figure 2.8.
Investments are restricted per period; this restriction reflects the players’ limited ability to add more production capacity (and also export capacity, see below). The data in Table 2.2 is based on historical
35 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
data and country reports, such as Eberhard (2015) and Lucarelli (2015). For all countries with several nodes, the value gives the aggregate of all producing nodes.
Another important parameter for a multi-period model is the discount rate applied to the profit functions of the producers and exporters. We use the costs of capital to determine the discount rate. The database of A. Damodaran at the New York University’s Stern Business School provides estimates of the costs of capital. In 2013 the data base reported 10.3% of the US coal industry. While this value is currently down to 5.7%, we believe that this is due to macroeconomic fluctuations, given the global value is at 9.0%.12 For convenience and to account for the additional uncertainty that is inherent in the market due to glooming climate policies we assume a 10% discount rate for exporters and producers.
Producers may, in theory, be limited by their available reserves, and we want to be able to capture possible incentives for producers to modify their behavior in reaction to constrained reserves. We follow the standard definition of reserves from the World Energy Council: “proved recoverable reserves are the tonnage within the proved amount in place that can be recovered (extracted from the earth in raw form) under present and expected local economic conditions with existing available technology” (EIA 2008).
Figure 2.8: Capacity and investment costs for all production nodes in the base year. Source: Authors’ work based on IEA and OECD (2008a) and Rademacher (2008).
12 See http://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/wacc.htm (accessed on July 27, 2016).
36 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
In Figure 2.9 we depict the distribution of global steam coal reserves by major global producing regions. In order to use consistent reserves data, we base ourselves primarily on one source (EIA 2008, table 8.2) but complement it with other sources for some individual nodes (e.g., DERA 2012). This data is aggregated on a national level; thus to get the distribution on a sub-national level other sources had to be used. For the US, EIA data (2008, table 15) was used; the reserve distribution for the Indian production nodes is based on GSI (2010) and that for the Chinese producers on NBSC (2007).
Table 2.2: Assumed production capacity expansion limitations per five-year period (in Mtpa). Production expansion Production expansion Country limitation (Mtpa) Country limitation (Mtpa) Russia 51 Poland 5 Ukraine 10 Kazakhstan 15 Venezuela 10 Colombia 30 China 450 Mongolia 20 Vietnam 10 Indonesia 90 Australia 40 India 110 Mozambique 15 South Africa 40 USA 224 Using a static reserves number is, of course, a limitation of our analysis, since real world observations show that the reserve assessments in the coal sector are sparse and often limited in geographical scope. Indeed, as a result of a large available reserve base and of costly exploration, the new delineation of resources that may potentially be mineable reserves is generally sluggish (Rogner et al. 2012). No structural analysis of resource to reserve conversion of the coal sector is available in the literature to our knowledge. However, since we do not want to neglect this fundamental aspect of production of a non-renewable resource, we opt for including regional reserve numbers.
37 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.9: Reserves of major countries in COALMOD-World (in Gt). Source: Authors’ work based on DERA (2012), EIA (2008), GSI (2010), and NBSC (2007). The coal quality data is shown in Table 2.3.13 The Energy Watch Group (2007) provides evidence that coal quality is generally decreasing over time as reserves are mined. According to this study the decline in coal quality is not only due to a shift toward lower rank coals, like sub-bituminous coals, but also to a quality decline within each class. The model captures some of this effect through the different coal qualities of the producers of the larger countries. For example, if the recent developments in the US continue with more (lower grade) coal from the Powder River Basin being produced, the overall quality of US coal will decrease.
Table 2.3: Energy content of coal by production node. Node Calorific value in kcal/kg Energy content in GJ/t USA PRB 4781 20.004 USA Rockies 6338 26.516 USA Illinois 6226 26.051 USA Appalachia 6949 29.075 Colombia 6375 26.673 Venezuela 6375 26.673 Poland 6300 26.359 Ukraine 6200 25.941 Kazakhstan 6000 25.104
13 Coal quality data is based on Platts (2009) for the US, Colombia, Venezuela, Poland, Russia, South Africa, Indonesia, and Australia. For China, it is based on Tewalt et al. (2010), and for India on Indian Bureau of Mines (2012, chapter titled “Coal and Lignite”). For Vietnam the quality data is taken from Ritschel and Schiffer (2007).
38 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Russia 6400 26.778 South Africa 6400 26.778 India North 4717 19.737 India Orissa 4187 17.520 India West 5209 21.793 India South 4866 20.360 Vietnam 7000 29.288 Indonesia 5450 22.803 China SIS 6597 27.600 China Northeast 5154 21.565 China HSA 6118 25.598 China YG 6074 25.413 Australia Queensland 6500 27.196 Australia New South Wales 6300 26.359 Mongolia 6100 25.522 Mozambique 6400 26.778 Source: Authors’ work based on Platts (2009), Indian Bureau of Mines (2012), Ritschel and Schiffer (2007), and Tewalt et al. (2010).
2.5.3 Land transport Land transport costs and capacities are associated with the transport from a producer to either a local demand node or to an exporter. In case of transportation to a demand node, the center of the demand region is used to calculate the distance; in the case of an exporter, the location of the main harbor is used. Land transport represents mainly transport by train but can also include road transport on trucks and in certain cases river transport by barges or overland conveyor belts. The transport costs are assumed to be constant over time and the capacities can be expanded by investments.
The transport costs for Colombia, Venezuela, South Africa, Indonesia, China, and Australia are based on Baruya (2007). For these countries, transport capacity data is based on relevant production, consumption, and export data. For the US, data for the transport costs is based on EIA (2011) for the transport costs. The transport capacities inside the US are determined using actual flow data given in EIA (EIA 2011, Spreadsheet “Domestic Distribution of U.S. Coal by Origin State, Consumer, Destination and Method of Transportation”). The land transport cost data for the CIS is from Crocker and Kovalchuk (2008) as well as IEA (2011) and the capacities are determined using relevant production, consumption, and export data. This method is also used to estimate the transport capacities in Vietnam and India. The Vietnamese costs were based on relevant price data. The Indian transport cost data is based on Indiastat (2012, Spreadsheet “Railway Freight on Coal in India”). Investments in additional overland transport capacity are set in a range between 10 and 55 USD/tpa depending on distance, landscape and topography, and if the project is mostly greenfield or not.
2.5.4 Export ports The data for the export ports includes the export capacity in the starting year and the port handling costs as well as the investment costs and investment limits per five-year period. For mathematical reasons, in the complementarity modeling format, we need to include a separate exporter for each producer in each exporting region. The coal of the dedicated exporter has the same calorific value as the producer and, thus, avoids the so-called “pooling problem” of mixing coals of different qualities. This way, we have, for example, an exporter “PRB” (Powder River Basin) next to an exporter
39 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
“Rockies” at the West Coast of the U.S. We needed to decide on the allocation of the total available capacities to these two export players. Similarly, where steam coal and coking coal exports use the same harbor facilities we had to decide on the capacity available to steam coal exports (e.g., in Australia). These decisions were taken case-by-case keeping in mind that capacities included in 2010 can be used at relatively low costs throughout the model’s time horizon without bearing investment costs for capacity expansions.
Cost data for most regions is based on Baruya (2007). For Colombia, the CIS, South Africa, and Mozambique, the capacity data in the starting year is taken from IEA (2011; 2012e). For Venezuela the costs are assumed to be similar to Colombia, and the capacities are determined using relevant export data. For the United States, the capacity data comes from VDKI (2012) as well as company information. VDKI (2011) provided information on Australia. Chinese port capacities are provided by the NBSC (2007). The costs for Poland are taken from Ritschel and Schiffer (2007) and the capacity is based on export data. Investment costs for additional export capacity are set between 10 and 30 USD/tpa depending on the country and the preexisting infrastructure. Data on allowed maximum expansion per five-year period is mostly from VDKI (2008; 2011) and IEA (2011). Figure 2.10 shows the unit costs of expanding export capacity together with the exporting harbor capacity in the base year.
Figure 2.10: Capacity and investment costs for all export nodes in the base year. Source: Authors’ work based on IEA and OECD (2008a) and Rademacher (2008). Constraints in export capacity expansion are particularly notable in global trade patterns. As above for the production investment limitations, the data in Table 2.4 is based on historical data and country
40 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
reports, such as Eberhard (2015) and Lucarelli (2015). For all countries with several nodes, the value gives the aggregate of all export ports. This means in particular that in the US (a country with four producing regions and three export ports, each possibly having several dedicated producers) the individual export limitation is lower. For example, the US West exporter has a total expansion possibility of 35 Mtpa per five-year period, but the PRB exporter via the West Coast has only 30 Mtpa investment capacity, the other 5 Mtpa of allowed investments on the West coast being allocated to the US Rockies producer.
China’s politically determined export restriction is assigned to the Chinese exporter. For 2010 we use a value close to the actual exports of 20 Mt. Forecasting the level of future export licenses is difficult and there are no such projections available. For the base case we assume the following values: 2015: 80 Mt; 2020: 90 Mt; 2025: 100 Mt; 2030, 110 Mt; 2035: 120 Mt; 130 Mt from 2040 onwards.
In sum, the cost of a ton of exported coal is the sum of production costs, land transport costs and the export fee. This is shown in Figure 2.11 for each exporter. In this figure, we also include the range of production costs in the respective production region. This cost range is represented by a white bar in the figure; it is calculated by subtracting the lowest average costs (black bar) from the highest average costs of the production region. (Positive shadow values of binding constraints on production, transportation or export constraints may also increase effective costs in the model but they depend on the market outcome and are therefore not depicted here.)
Table 2.4: Assumed export capacity expansion limitations per five-year period (in Mtpa). Export expansion Export expansion Country limitation (Mtpa) Country limitation (Mtpa) Russia 65 Poland 5 Ukraine 10 USA 115 Venezuela 10 Colombia 40 China 50 Australia 108 Indonesia 50 Mozambique 15 South Africa 10
41 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.11: FOB costs (2010) for the export countries in COALMOD-World (in USD/t). Source: Authors’ work based on IEA (2012e); IEA (2011); Baruya (2007); NBSC (2007); and Ritschel and Schiffer (2007).
2.5.5 Freight rates Overseas shipment is a cost to the coal importers that we approximate by the freight rates paid to the shipping companies. Freight rates result from the supply-demand equilibrium in the dry bulk carrier market and their quotations have been very volatile in the past.14 In general, the freight market behaves cyclically. This makes it difficult to predict future freight rates, which are needed as a transportation cost input for the model. Moreover, for the same route there is a difference between Capesize and more expensive Panamax freight rates; the capacity of Capesize ships is higher but Panamax vessels are used more often on shorter routes. In the model, we assume the freight rate (transport cost) to be dependent on distance to reflect the spatial character of the international coal market.
14 Dry bulks include commodities such as iron ore, coal, or grain.
42 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.12: Linear regression of average freight rates between 2002 and 2009 (in USD/t). Source: Authors’ work based on Platts newsletters 2002–2009. Given historical information on weekly freight rates on all available routes, we specify a linear regression using distance as the explanatory variable. This is necessary because freight rates are only reported for the major shipping routes and not for all exporter-importer pairs that we include in the model. We specify a different regression for 2010, which is based on 2010 observed data (obtained from Clarksons for all available coal shipping routes, including Panamax and Capesize), than for the subsequent model periods 2015–2040. For all future periods, we use data from 2002 to 2009 as input to the regression, to obtain long-run average costs.15
The regression equation for freight rate as a function of distance, determined using observed freight rate averages between 2002 and 2009 is y = 0.009x + 12.419. The computed values for y are used as shipping costs and are set constant from 2015 until 2040. The shipping costs between every export node and every import node with import possibility are calculated using this equation by plugging in the corresponding distance x.16 Table 2.5 gives calculated freight rates for some main routes – for example, from South Africa (Richards Bay) to Northern Europe (Rotterdam). Freight rates on each route, in addition to export port fees, are added to the FOB costs depicted in Figure 2.11, which gives the CIF costs as shown in Figure 2.13. This figure does not include possible additional shadow values of restricted capacities (of production or export capacities) that may be computed in the model runs.
15 Sources for weekly freight rates (end-of-week quotations) are Platts newsletters from 07/2002 to 10/2009 (where extreme values have been removed from the data). 16 Distance is calculated using the PortWorld online distance calculator; see http://www.portworld.com/map/.
43 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Table 2.5: Freight rates for selected routes (in USD/t). From Australia - Queensland To 2010 2015-2030 Rotterdam 20.47 23.72 Japan 14.97 15.69 From Australia - New South Wales To 2010 2015-2030 Rotterdam 20.33 23.52 Korea 15.38 16.42 From Colombia - Puerto Bolivar To 2010 2015-2030 Rotterdam 15.44 16.5 From South Africa - Richards Bay To 2010 2015-2030 Rotterdam 16.94 18.65 Chennai 15.09 16.01 From US West - Portland, OR To 2010 2015-2030 Rotterdam 18.08 20.29 Shanghai 15.72 16.91 From Indonesia - Banjarmasin To 2010 2015-2030 Rotterdam 19.84 22.82 Guangzhou 13.76 14.09
Figure 2.13: CIF costs in 2010 for selected routes (in USD/t).
44 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
2.5.6 Demand – Two possible scenarios For the specification of the demand function of each consumption node, we need the “reference price” and “reference quantity” point for each model year. To obtain a consistent demand database for all countries in the model we use data from the International Energy Agency (IEA 2012b; IEA 2012c; IEA 2015a). IEA (2015a) reports on two scenarios which we use and contrast in the model: the New policies scenario in which coal demand continues to grow but at a much slower pace totaling in a 12% over the period 2013 to 2040; and the 450ppm scenario in which coal demand is significantly reduced almost to levels of 2000 (a reduction of 36% in the period from 2013 to 2040) due to some additional climate policy efforts.
The IEA (2012b; 2012c) reports demand data for 2010 of each consumer in Mtpa. This was converted to Petajoules using the calorific values for the main supplier from Table 2.3. IEA (2015a) gives coal demand projections until 2040 in the two scenarios. This is regionally aggregated data and steam coal is not distinguished from coking coal or lignite. Hence, for our demand scenarios we apply the regional growth rates between 2010 and each reported later period to the 2010 demand data obtained from IEA (2012b; 2012c) instead of using the reported consumption values directly. We generally use the growth rate of the reported coal consumption for power generation, except for China, India, and South American countries where we use the growth rate of the TPED (total primary energy demand) for coal.17 This is due to the fact that China and India, but also Chile and Brazil currently still use a lot of steam coal in other consumption sectors (households, industry) and that, while we have to include this consumption, we need to take into account its falling tendency as reflected in the projected TPED growth rates.
Price data was taken from the IEA (2012b, CIF prices and steam coal prices for electricity generation). The prices in 2010 to 2050 in each scenario were calibrated such as to obtain the reference consumption in all demand nodes (with a margin of 1% or less) based on a cost-minimization approach for supply. The calibration resulted in an average price increase between 2010 and 2040 of 1.2% p.a. in the Stagnation scenario, which is derived from the New Policies scenario and reduction of 0.35% p.a. in the 2°C scenario, which is derived from the 450ppm scenario. These tendencies are in line with the trends projected by IEA (2015a).
Own-price elasticities of coal demand are part of our demand curve definition. However, empirical research on elasticities, especially for coal, is scarce and the results are often not very satisfying. Dahl (1993) estimates short run elasticities of coal between -0.55 and -0.3. Aune et al. (2004) use a value of -0.19 for the short run elasticity of coal demand in their model. Liu (2004) yields a rather peculiar result of a zero elasticity that is of rather limited use for defining demand functions for the model. We conclude that the price elasticity of coal demand is rather inelastic and assign elasticity values of -0.1, -0.2, or -0.3 to the model consumers, based on the percentage of coal use in the total power generation: the more dependent a country is on steam coal use in its electricity sector, the less elastic its demand is assumed to be. We take into account a higher long-run elasticity of demand, which
17 In this, our approach is consistent with IEA (2015a, chapter “Coal Market Outlook”).
45 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
allows a structural change in the demand for coal depending on the expected future prices, by gradually increasing the price elasticities of each period over time: the values reached -0.4 in all countries by 2020, -0.5 by 2025, and -0.6 by 2030 and subsequent years.
All consuming countries that are modeled as importers without domestic production are only included with their import demand by subtracting the respective local production from demand in the 2010 data. Where 2010 demand has to be allocated to several consumption nodes in one country, country- specific information is used.18
Table 2.6: Reference consumption in 2020, 2030, 2040, by IEA region from New Policies and 450ppm Scenario, and extrapolation for 2050 (in % of 2013 consumption).19 New Policies Scenario 450ppm Scenario Region [%] 2020 2030 2040 2050 2020 2030 2040 2050 Data origin WORLD 103 107 112 112 95 74 64 45 TPED coal OECD Americas 85 68 60 45 69 34 38 14 Coal in Electricity USA 84 69 61 46 68 34 39 17 Coal in Electricity OECD Europe 84 52 33 2 74 25 18 5 Coal in Electricity European Union 82 46 27 4 73 25 19 5 Coal in Electricity OECD Asia Oceania 90 79 65 53 84 38 16 4 Coal in Electricity Japan 89 83 71 66 81 34 7 3 Coal in Electricity E. Europe/Eurasia 93 87 87 84 84 44 29 5 Coal in Electricity Russia 96 99 96 96 84 44 33 1 Coal in Electricity Non OECD Asia 113 131 148 148 104 72 53 25 Coal in Electricity China 100 101 96 96 96 75 61 42 TPED coal India 140 202 274 274 130 135 133 133 TPED coal Africa 111 127 152 152 88 81 66 56 Coal in Electricity South Africa 99 95 89 84 96 75 59 40 TPED coal Latin America 121 158 192 192 117 104 108 100 TPED coal Brazil 125 144 163 163 119 100 94 81 TPED coal
2.6 Modeling results until 2040
2.6.1 Scenario assumptions: stagnating coal demand or climate policies with significant demand reduction For each model year, the COALMOD-World model delivers results for the inland and seaborne trade flows, the prices, the level of investments, and the value of the dual variables of the constraints that indicate if and how strongly a specific constraint is binding. The results for the last two model years 2045 and 2050 are not presented as there is a risk of distortion because there is less incentive to
18 For the US, data from the EIA (2010, Spreadsheet Domestic Distribution of U.S. Coal by Origin State, Consumer, Destination and Method of Transportation) is used. For Russia, the allocation of demand is based on the regional location of coal-fired power plants given by EFA (2008). For China, the coal flows reported in Mou and Li (2012) are used to obtain the regional breakdown of demand. For India, Indiastat (2012, Spreadsheet Region/State-wise Linkage, Receipt, Import and Consumption of Coal in Various Thermal Power Stations) gives regional information. 19 IEA (2015a) only covers the period until 2040, and only reports values for 2020, 2030, and 2040. Values for 2025 and 2035 where obtained by linear interpolation from the adjacent periods. Values for 2045 and 2050 were obtained by linear interpolation from value from 2020 to 2040. If the interpolation would have yield an increase we assumed constant demand from 2040 onwards, if the interpolation would have yield negative demand we assume a reduction of the 2040 by 50% for 2045 and by 75% by 2050.
46 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
invest without any possible revenue after 2050. For convenience, we only present the results for the years 2010, 2020, 2030, and 2040 here.
Our results are based on the assumption of competitive and liberalized markets.20 We also assume that the markets are fully integrated, that is, when a demand node can be reached by different producers or can import coal from overseas, it can fully substitute between the different sources. We assume that no fundamental structural change in the coal market will happen during the model horizon. The model results can be called “ideal” results, as they tell us how future demand should be served optimally and in which countries investments should take place. We further assume that there are no policy-based restrictions on trade apart from the case of Chinese exports in 2010.
We examine two scenarios: the “Stagnation scenario”, which is derived from the IEA’s World Energy Outlook’s New Policies scenario, and the “2°C scenario”, which is derived from the IEA’s World Energy Outlook’s 450ppm scenario (IEA 2015a).
The two scenarios fundamentally differ in the energy mix induced by climate policy efforts and, hence, in CO2 emissions. Globally and from all sectors, annual energy-related CO2 emissions increase by 16% in the New Policies scenario between 2013 and 2040. By contrast, in the 450ppm scenario global emissions are reduced by 41%. The relative contribution of coal to total CO2 emissions by 2040 is forecast to be slightly reduced from 46% to 42% in the New Policies scenario and but substantially decrease to 24% in the 450pmm scenario.21 In other words, at the global scale, a structural transformation involving a drastic shift away from coal is expected in the second scenario.
Regional consumption patterns are affected in the two scenarios by the instruments and relative stringency of climate policy across regions. The IEA scenarios assume the implementation of the 2030 Climate and Energy Package for the EU, a phase-out of fossil fuel subsidies for all Non-OECD countries where such policies were already announced, the implementation of various policies directed at reducing fossil fuel consumption and improving air quality and also some mitigation efforts in India. To achieve stringent emission targets in the 450ppm scenario the storyline assumes introduction of
CO2-pricing regime by 2020 for all OECD and major Non-OECD countries, and a comprehensive phase-out of fossil fuel subsidies in Non-OECD countries.
We incorporate these national differences in the scenario assumptions in COALMOD-World by using the coal consumption projections from the respective IEA scenarios. Specifically, we use the demand growth projections for coal for power generation in the respective scenarios, with the exception of China and India, South Africa, and Latin American countries where the growth rates of Total Primary
20 We are aware that not all countries currently have fully liberalized domestic markets (e.g., India). However, we assume that the markets’ structure or outcomes will move toward competitiveness in the future. 21 CCTS plays at best a marginal role in reducing CO2 emissions from coal consumption in the New Policies scenario with 3% of the coal-fired generation fleet being equipped with Carbon Capture technology. By contrast, the 450ppm scenario relies heavily on CCTS with a penetration of 75% by 2040. We take a critical view on the CCTS technology assuming that even strong climate change mitigation efforts will not lead to a significant deployment (see Hirschhausen et al. (2012a) as 0 of this dissertation for perspectives on the prospects for CCTS technology). Therefore, we interpret the coal demand patterns implied by the 450ppm scenario as an upper bound, with even stronger reductions of coal consumption necessary in the absence of CCTS.
47 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Energy Demand (TPED) from coal are used (for more details, see section 2.5.6 and, in the Appendix, Table A.5). This gives the global steam coal demand for each scenario as shown in Figure 2.14.
We assume that there are restrictions on export capacity expansion for technical and economic reasons. These restrictions are based on historical experience as well as on planned and forecasted expansions. They range from 5 to 30 Mtpa of additional capacity that can be added over a five-year period depending on the country. We include export restriction for US coal via west coast ports in line with the current lack of such domestic capacity (minor amounts are exported via British Colombia) and persistent concerns about environmental and health impacts (Western Interstate Energy Board 2015). We do not impose restrictions on expansions of inland transport capacity due to a lack of detailed data on this for all modeled countries.
Figure 2.14: COALMOD-World results: development of yearly global coal demand in both scenarios until 2040 (in Mtpa).22
2.6.2 Overview of results: stifle Asian “hunger” for coal Figure 2.15 and Figure 2.16 give an overview of the evolution of consumption in the world steam coal market in our Stagnation scenario and 2°C scenario and provide more detail on the supply structure. The total surface of this graph represents total consumption, and the different areas decompose the consumption by its origin, seaborne trade or local domestic supply with a special focus on India and China. In the Stagnation scenario, global coal consumption will remain more or less constant with a moderate increase of 5% from 5000 Mtpa in 2010 to 5235 Mtpa in 2040. Most increase in demand
22 COALMOD-World is an energy-based model that calculates trade flows in Petajoules. For better representation, the results shown are aggregated and expressed in Mtpa. These values are calculated using the relevant quality factors of each producer. Detailed results are reported in Table A.8 in the Appendix.
48 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
originates from rising consumption in India. Rising demand in South-East Asian countries is overridden by decrease in demand from OECD countries.
.
Figure 2.15: Global COALMOD-World results: aggregated consumption and imports in the Stagnation scenario (in Mtpa).
Figure 2.16: Global COALMOD-World results: aggregated consumption and imports in the 2°C scenario (in Mtpa).
49 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
In contrast, the 2°C scenario sees a decrease in steam coal consumption by more than 40% to 2091 Mtpa in 2040. All countries drastically reduce their demand with a steeper decease between 2020 and 2030 and a less steep reduction before and after. The stark exception is India, where consumption increases by 29% until 2040. This corresponds to an average emissions reduction of 3.6 GtCO2/a over the model horizon.
Some similar trends can be observed in both scenarios. Around 20% of coal consumed is imported traded on the seaborne market in 2040, up from about 15% in 2010. In both scenarios, the largest share of consumption takes place in Asia, in particular in China and India. Asia23 accounts for about 68% of global consumption in 2010, this share increases to 79% until for both scenarios, with China alone consuming about 45% (Stagnation scenario) to 51% of the global coal supplies by 2040.
As China’s import behavior is rather driven by arbitrage that by structural supply restrictions (e.g., Morse and He 2015), it is not surprising that the share of imports in total consumption is higher in the 2°C scenario (19% as compared to 7% in the Stagnation scenario). Chinese coastal consumers benefit from lower market prices which originate from overall lower demand. In general, India relies more heavily on imported steam coal with its share rising from 11% in 2010 to 33% in 2040 for the Stagnation scenario and more moderate 24% in the 2°C scenario. In absolute terms this corresponds to a 770% increase from 2010 to 2040 for the former and a 282% increase for the latter scenario. In 2010, the two countries account for 23% of seaborne trade. Already by 2025, their imports account from more than half of seaborne trade. This share rises to 60% and 75% until 2040 for the Stagnation scenario and the 2°C scenario, respectively.
Table A.7 in the Appendix shows that, in addition to the big players China and India, other Asian countries also increase their coal demand to a varying extent in the two scenarios. Indonesia, to date mostly known as one of the largest exporters, is expected to see a moderate growth of domestic coal consumption – a development which may also influence its exporting behavior if new mining areas are too costly to be added to the reserve base. Moreover, other South-East Asian importers –Taiwan, Thailand, Malaysia, and the Philippines – will increase their demand in the Stagnation scenario. Besides, Latin American consumers like Chile, Mexico, and Brazil are also expected to increase consumption.
2.6.3 Global trade results
2.6.3.1 A shift to Asia The major focus of the modeling effort is on investigating international trade flows. We obtain long run equilibrium results which abstract from the volatility of prices or costs, which may cause coal trade flows to diverge from real-world observations in the short run. Our modeling approach implicitly assumes that the production and investment quantities as well as the trade flows of the equilibrium path will be reflective of long-run trends.
23 Without Israel, Turkey, Central Asia and Russia, and Oceania.
50 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
The results for 2010 show a notable similarity with the actual observed trade pattern. This is an important achievement, given that we not only simulate the trade flows shown on the maps in Figure 2.17 to Figure 2.20 but also internal markets. We have a global integrated market with flows from South Africa and North America, traditional suppliers of the Atlantic basin, to Asia in the Pacific basin. The direction and relative amounts of the trade flows are a reasonable match to actual trade flows.
Figure 2.17: Global results 2010: seaborne trade flows (in Mtpa). Figure 2.18 and Figure 2.20 show the trade flows in 2020 and 2040 in a stylized and simplified representation, distinguishing Stagnation scenario and 2°C scenario results. Most obviously, the demand reduction in Europe and North America and the consumption increase in Asia lead to a noticeable shift of international trade flows to Asia. This is the case not only in the Stagnation scenario but also in the 2°C scenario. This is due to the concentration of climate and energy efficiency policies in Europe and North America, as well as China and OECD Asia, while South-East Asia and India pursue a carbon-intensive development path. The model predicts that the overall picture for the global market will have significantly changed by 2040: Russia and Poland are the only suppliers to Europe; South Africa, Europe’s traditional supplier, will have become a major supplier to India and the Pacific market; and Colombia also mainly delivers to the Pacific market.
Australia and Indonesia have traditionally been the key players in the Pacific market. While Australian exports increase by 28% until 2040, Indonesian exports stagnate until 2030 and then drop by 12%. This is due to a limited resource bases, and an intertemporal optimized extraction path.24 Nevertheless, Indonesia will consolidate its role as the leading steam coal exporter, ahead of US, South Africa and Colombia (cf. Table 2.7). Relatively low production costs and flexible, low cost investments are the main reasons for this development.
24 In our model results, the Indonesian reserves constraint is binding until 2050, in the New Policy scenario. In other words, all steam coal currently known to be mineable in Indonesia will be exploited until 2050, and even more will be produced if the reserve base increases. One can reasonably extrapolate from past observations that more exploration activity and an increase of the reserve base will take place in Indonesia once the remaining reserves have passed below a certain threshold.
51 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.18: Global results 2020: seaborne trade flows in both scenarios (in Mtpa).
Figure 2.19: Global results 2030: seaborne trade flows in both scenarios (in Mtpa). In some circumstances the US will join the group of large exporters. Challenged by a decline in consumption due to the availability of cheap shale gas, and by strict environmental and climate regulation (most notably the Clean Power Plan) the US coal industry is in crisis with major bankruptcies in 2015 and 2016 (EIA 2015b).25 Sustaining the trend, coal from Powder River Basin (PRB) in Wyoming and Montana can extend its dominate role on the US domestic market, due to its cost advantage and abundance. Given that there will be no climate policies addressing leakage from US domestic coal markets to international markets, the latter may be an outlet to the stranded coal if international prices recover. In our model high cost US suppliers increasingly export coal to Asian, and to a smaller extent to Latin American markets. Until 2040, the US becomes the second largest supply to international markets, up from the 8th rank in 2010.
25 Mooney and Mufson (2016): “How Coal Titan Peabody, the World’s Largest, Fell into Bankruptcy.” The Washington Post, April 13. https://www.washingtonpost.com/news/energy-environment/wp/2016/04/13/coal-titan- peabody-energy-files-for-bankruptcy/ [accessed 23.07.2016, 15:41].
52 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.20: Global results 2040: seaborne trade flows in both scenarios (in Mtpa). The third most important exporter in 2010 was South Africa. Given further increase of its exports after an expansion of its critical railway capacities to the ports, South Africa is predicted to retain its ranking. After years as swing supplies between Europe and Asia, we expect South African exports quickly to be redirected to Asia alone, and especially to India. We can see the emergence of a third market (in addition to the traditional Atlantic and Pacific markets) that could be called the “Indian market”, where South Africa will become the key player. However, the depletion of current operating mines and lower quality and long transport distances for new mines might reduce South Africa’s ability to dominate this market, especial if India’s plans to shift to modern coal-fired power plants which required high quality coal are realized (Eberhard 2015; Commonwealth of Australia 2015).
2.6.3.2 Other trends Colombia, and with a similar pattern the smaller Latin American exporter Venezuela, has traditionally been a supplier to the Atlantic basin. In 2010, Europe was its main export destination. However, through the extension of the Panama Canal, these suppliers now also have easier access to the Asian market.26 Due to low demand and low prices, there is a major swing to the Asian markets until 2020, and a full reorientation until 2030, according to the simulation. Colombia has repeatedly faced infrastructure bottlenecks in the past (e.g., in its railroad transportation). If these are resolved, Colombia can considerably increase its market share thanks to its relatively low production costs.
Russia strengthens its position as an important supplier in the European market, taking over market shares from South Africa and Colombia, which turn more to the Asian market. By 2030, approximately 86% of European steam coal demand is served by Russia in both scenarios. Against the background of concerns about security of supply from Russia this might not be a favorable development for European consumers. While the model forecasts a price differential of about 8% between Europe and prices paid in India and China, this gap might be smaller if the Europeans opt for a self-imposed
26 Wallis(2016): “Expanded Panama Canal: Bigger Ships, Bigger Paydays for Beans, Coal, Gas.” Reuters. June 25. http://www.reuters.com/article/us-panama-canal-commodities-idUSKCN0ZB0Z0. [accessed 24.07.2016, 20:57]
53 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
diversification of suppliers. In both scenarios, Russia is also increasingly active on the Asian market with supplies to Japan, Korea and China. To a smaller extent it is also active on the Latin American market.
In the model scenarios, China will cease its exports to other countries due to its strong domestic demand, and diminishing cost advantages compared to Indonesian suppliers. Hence, the export restriction, which was still binding in the last decade, will play no role any more. China increases its import levels in the next decades on different paths in the two scenarios: in the Stagnation scenario, imports gradually increase from just under 100 Mtpa in 2010 to 278 Mtpa in 2030; and to 313 Mtpa in the 2°C scenario. In both scenarios there is a drop after 2030, which is more pronounced in the Stagnation scenario (down to 168 Mtpa) and less pronounced in the 2°C scenario (down to 286 Mtpa). The divergence is driven by lower global demand which results in increased availability of low-cost coal on the international market.
Table 2.7: Share and rank in international trade flows of major exporters in both scenarios and over time. Stagnation scenario 2°C scenario Share of Rank in Share of international Rank in international international trade international trade [%] trade [%] trade 2010 2020 2030 2040 2010 2020 2030 2040 2020 2030 2040 2020 2030 2040 AUS 13 12 11 11 2 4 5 5 12 11 10 3 3 4 COL 9 12 13 13 5 2 3 4 9 8 7 5 5 5 IDN 39 32 27 22 1 1 1 1 34 38 39 1 1 1 MNG 2 5 4 1 10 7 7 10 4 4 4 7 6 6 RUS 10 11 10 10 4 5 6 6 11 11 11 4 4 3 USA 3 5 15 19 8 6 2 2 3 1 1 10 9 9 ZAF 10 12 12 14 3 3 4 3 13 18 22 2 2 2 Interestingly, China also will rely on coal supplies from Mongolia in addition to imports from the seaborne market that can land in the demand regions along the coast. Mongolia has recently started to scale up its coal production and will continue to do so according to the model by relying on the stable market for its coal in China.
The supply relations in Asia also explain the expansions of export capacities of the major suppliers and their sensitivity to demand changes. Table 2.8 reports the difference of export capacity expansions in 2°C scenario compared to the Stagnation scenario for some major exporters. As could be expected from the assumption of policies that curtail coal demand after 2020 in all countries except India, there is no need for export capacity expansions for any of the established international suppliers. It is only the new entrants like Venezuela and Mozambique who expand their capacity after 2020. The exception is South Africa, which consolidates its position on the “Indian market”, and substantially expands its export capacity, though still by less than in the Stagnation Scenario.
The reduced demand in the 2°C scenario is, of course, reflected in slower production growth. Less new production capacity has to be built to accommodate the lower consumption levels, as reported in Table 2.8, which shows the difference between production capacity expansions in the 2°C scenario
54 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
compared to the Stagnation scenario.27 Most notably, the US and Colombia, and to a smaller extend South Africa, exhibit significantly reduced export-oriented production capacities in the 2°C scenario.
Table 2.8: Export capacity and production capacity: results of 2°C scenario compared to Stagnation scenario (in Mtpa). Export Producer 2020 2030 2040 2020 2030 2040 country country AUS 0 0 0 AUS 0 0 -7 COL -42 -66 -69 CHN -20 -324 -362 IDN -12 -20 -20 COL -42 -77 -96 MOZ -1 -1 -1 IDN -18 -49 -46 RUS -11 -15 -17 IND -15 -162 -349 UKR -5 -5 -5 MNG -8 -13 -11 USA -35 -75 -127 RUS 0 -2 -14 VEN 0 -4 -4 USA -59 -91 -94 ZAF 0 0 -10 ZAF -10 -33 -54
2.6.4 Price analysis COALMOD-World calculates prices in each demand region. The price is the fundamental signal that attracts coal into a market and reflects that market’s willingness to pay. In our perfect competition setup, the price equals the production and transportation costs of the highest cost supplier (the so- called marginal supplier, i.e., the supplier of the last, marginal unit) plus any possible shadow prices of the various constraints. Hence, prices may vary significantly between regions depending on the willingness to pay (demand function) and the availability of high-cost and low-cost suppliers. Regional (nodal) prices are calculated in the model calibration mechanism so as to obtain the consumption levels set by the reference data. In other words, the price results show what price level is necessary to reach a given consumption level. Our model cannot predict short-term price volatility but gives long- term price trends based on the fundamentals of the market.28
Figure 2.21 shows the price evolution for some major regions in the Stagnation scenario. Globally, prices show an upward trend over the time period 2010 to 2040. We observe continuous price increases of 1.22% p.a. on average in the Stagnation scenario and a reduction 0.4 % p.a. in the 2°C scenario until 2040 – that is, an increase of 37% or a decrease of 10%, respectively, over the time horizon.
The lowest prices are the domestic prices in the US, South Africa, and Russia. These demand nodes are close to large and cheap sources of supply and are not connected by imports to the global market; thus, they are less affected by higher prices in other regions. While the latter two still show some moderate increase over time, domestic US prices decrease from 54 USD/t to 52 USD/t, with a low of 49 USD/t in 2020. This is in line with the observation of reduced demand, couple with overcapacity and cheap supply from Powder River Basin.
27 Note that this table reports on the levels of production capacities and does include absent investments in new mines that replace old mining capacities in line with the mining mortality mechanism. 28 Short-term volatility can be caused by extreme weather events, strikes, infrastructure problems, conflicts etc. on the supply side. On the demand side, we can consider shocks that come from the energy system such as short- term problems with nuclear reactors or low water level of hydro-power plants.
55 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.21: Average prices for selected regions for all model years (in USD/t) in the Stagnation scenario. In the near term, highest prices are the import prices in Europe, the Mediterranean countries and in particular in China due to the long transport distances (with high transport costs) and the high willingness to pay. The second highest increase over time can be seen in the South-East Asian prices. This is due a rising demand that is met by supply where the marginal supplier (often Australia) has faces increasing production costs. The highest increase, however, can be observed with Indian import, and even more so with Indian domestic prices. With no access to the international markets and a strong demand cost of supply and therefore also prices increase more than double from 2010 to 2040. Two opposing effects govern the trend of European prices. On the one hand low cost suppliers increasingly turn to Asia to ear there margins, therefore consumers have to rely on high cost supply from Russia and Poland. On the other hand, when demand is drastically reduced especially in 2030 and onwards, prices decrease again.
It is noteworthy to look at the development of production costs in Figure 2.22 and Figure 2.23 since the prices reflect the costs along the value chain of the marginal supplier in the perfectly competitive market (in addition to shadow values of binding capacity constraints). The production cost function changes over time (i.e., shifts upward) due to the mine mortality mechanism in the model, which leads to an upward movement of the cost curve’s intercept. The upward trend is more or less fast for each producer, depending the respective mine mortality, but also on additional production capacity expansion. For example it is considerably slower for the Powder River Basin than for Colombia or Australia. The decisive point of the cost curve is the cost level of the last produced unit; this is what is depicted in its development over time in Figure 2.22 and Figure 2.23 (compare to the 2010 cost curves in Figure 2.7 in section 2.5.2).
56 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
Figure 2.22: Production costs at production level for selected producers over time in the Stagnation Scenario.
Figure 2.23: Production costs at production level for selected producers over time in the Moderate Growth scenario. Both figures confirm the picture of the base year (2010) cost order for the following years: Australia, and the US Appalachian are the most expensive producer, followed by China SIS. South Africa, Colombia, and Indonesia are in the group of medium cost suppliers, undermined by Russian coal,
57 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
which faces long-distance haulage, though. The US Powder River Basin is the cheapest supplier on the world market. The grouping remains more or less unchanged throughout the model horizon, with the exception of Colombia who’s production costs rises faster than those of South African and Indonesian competitors.
The lower consumption levels of steam coal in the 2°C scenario lead to a lower cost (and price) level than in the Stagnation scenario. The decrease in production cost which can be observed in Figure 2.23 originates from a decrease in production levels. Most pronounced cost reductions can be observed for the Appalachian region, Australian and Chinese producers. While highest costs are found at 82 USD/t in Queensland, Australia in 2040 in the Stagnation scenario, costs are down to 56 USD/t in the same region in the 2°C scenario.
2.7 Model limitations While COALMOD-World constitutes a data-rich and comprehensive modeling framework that can be applied for various types of analysis, the model still has some inherent limitations that are briefly discussed in this section.
In the model demand response and induced changes in supply patterns are driven by price elasticities. The framework assumes an inverse demand function which is estimated based on reference demand, reference prices, and some ideal about the point elasticity. In contrast to a General Equilibrium framework, the commodity that is used to substitute reduced demand in coal is not specified. Therefore, calculated emission reductions for any policy scenario must be considered as upper bounds, as the substitution may also be towards other fossil fuels, e.g., natural gas rather than to zero emission renewable energy sources.
Moreover, the model does not take into account macro-economic parameters and Interaction of coal prices with other key commodity prices, most notably the oil price, as a major input to production costs as fuel for machinery and transport. Exchange rates, which are commonly adjusted by exporting countries to mitigate price fluctuations, are also not considered. Such interactions, just like industry business cycle effects, have strong short-term implications on coal prices. As COALMOD-World is more concerned with the medium to long-term development of the steam coal market, the model abstracts from such short-term adjustments. For the analysis of supply and demand-side disruptions or market power exertion such additional determinates would highly increase the accuracy of results.
It is also worth noting that coal quality in terms of gross calorific value is an important characteristic of coal that determines the value of the commodity but also its substitutability. Often, power plants are tailored to a particular type of coal that is required for the combustion process and deviations induce additional costs due to reduction in efficiency and additional maintenance. The model does not account for this fact, and assumes that all steam coals are perfect substitutes based on their energy content. Moreover, the model assumes that quality of coal for a particular producer is constant over time. In reality, the quality of coal produced from a mine can vary substantially over time depending on geological characteristics of the current coal seam. Haftendorn, Holz, and Hirschhausen (2012)
58 Chapter 2: The COALMOD-World Model: Coal Markets until 2040
propose a formulation of COALMOD-World that accounts of changing quality over time. Due to a lack of data for parametrization the formulation presented in this chapter does not include this feature.
Finally, the model focuses on steam coal which is relevant for the international market. There are interrelations and some substitution between other types of coal, i.e. lignite and coking coal. There is a fluent passage between lignite and steam coal and also between steam coal and coking coal, where the latter are often mined in the same mines. Policies directed towards the one commodity might well have effect on the other. A comprehensive approach covering all types of coal and their particular use would deliver a more holistic picture but is difficult to accomplish due to the heterogeneity of cases.
2.8 Conclusions In this report, we have presented a tool for examining the future global steam coal market, the COALMOD-World model. We have shown how we can model this market and its future development using a large-scale equilibrium model that relies on microeconomics and game theory. The combination of model theory and detailed market analysis provides the ground for the development and the implementation of the model.
The illustrative results are based on two IEA World Energy Outlook scenarios (IEA 2015a): The market equilibria obtained from demand projections based on the “New Policies Scenario”, which includes some minor implementation of climate policy resulting in a stagnation of coal demand, and the “2°C scenario” which induces a coal demand path consistent with the 2°C target and is derived from the IEA 450ppm scenario.
While both scenarios share the general trend of a shift of the international steam coal market towards Asia, they imply two fundamentally different development paths for the steam coal market. Demanded quantities differ by over 46% in 2040 and observed prices vary between an almost 40% increase and a 10% decrease. Internationally traded volumes diverge by almost 500 Mtpa in 2040 with different trade relations, and required investments in production, transport and export capacity implied by the two scenarios.
The comparison demonstrates the functionalities of the model and provides examples for possible insights gained from the modeling exercise. At the same time, the discussion also illustrates that the model results can only be interpreted in the context of a specific market situation and political context, in combination with idiosyncratic expertise of the modelers
59 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
Chapter 3 TESTING SUPPLY-SIDE CLIMATE POLICIES FOR THE GLOBAL STEAM COAL MARKET – CAN THEY CURB COAL CONSUMPTION?
3.1 Introduction: Supply-side climate policies as an alternative route to achieve desired emission reductions The COP21 Paris agreement has brought about a clear commitment to reduce anthropogenic greenhouse gas (GHG) emissions to a level that will most likely keep the increase of global mean temperature below 2°C29 and striving for 1.5°C.30 McGlade and Ekins (2015) estimate that achieving the 2°C target requires refraining from using a large share of current fossil fuel reserves but leaving them in the ground. Given its limited use for other than heat generation and resulting low economic value (Collier and Venables 2014) on the one hand and its abundance on the other hand, 82%-88% of current coal reserves need to be left unburned until 2050 (McGlade and Ekins 2015). The difference in the two numbers accounts for possible future use of Carbon Capture, Transport and Storage (CCTS), a technology which is currently not available at a demonstration scale31 and which has thus far not lived up to the high hopes put in it (Reiner 2016).
While there is consensus that reducing CO2 emissions and refraining from coal consumption are inseparably linked, there is major inertia hindering the transformation of the energy system. Incumbent industries in countries that have a long history of using coal as the primary fuel in their energy mix are reluctant to adapt their business models and to bring forward decarbonization (Fulton, Spedding, et al. 2015). Although a large number of demand-side policy instruments exist (see section 1.2.4.1) they are not sufficient to achieve required emission reductions. In fact, the IEA World Energy Outlook New Policies Scenario (IEA 2015a) which assumes the implementation of most of currently announced climate policies, including most of the Intended Nationally Determined Contributions (INDCs)32 under
29 Hereafter referred to as the “2°C target”. 30 This chapter is a single author publication submitted to Climatic Change. Furthermore, was published as a DIW Berlin Discussion Paper No. 1604, 09/2016 (Mendelevitch 2016b). 31 The only existing CCTS infrastructure at Boundary Dam in Saskatchewan, Canada (in operation since October 2014), uses the CO2 for enhancing oil recovery and thus cannot be considered an emissions reducing project. See Part B for more details on CCTS. 32 With one major exemption: INDCs submitted by India are not fully incorporated but rather the original target of 100 GW of solar PV installed until 2022 is reduced to 40 GW (IEA 2015a, 498).
60 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
the United Nations Framework Convention on Climate Change (UNFCCC), still projects a 15% increase of annual global emissions until 2040. Coal production is expected to increases by 18% during the same period. Even though the scenario fails to incorporate some of the major trends with respect to the restructuring of global energy systems33, the general conclusion that currently discussed policies will not lead to a deep decarbonization, is still valid.
While most of these policies are directed towards the demand-side of fossil fuels, many scholars argue that supply-side policies hold promise to be more effective in achieving desired emission reductions (see e.g., Lazarus, Erickson, and Tempest 2015). The contribution of this chapter is to quantify the effects of two supply-side policies that are currently discussed to complement the wide range of demand-side policies in further reducing fossil fuel consumption: The first instrument is a removal of coal production subsidies to reveal the “real” cost of coal supply. This policy measure can be seen as one part of the international strive to phase out fossil fuel subsidies, as agreed on, e.g., by the G20 (2009). This chapter contributes to the literature by summarizing available information on coal production subsidies in the major producing countries and providing an estimate on the mark-up resulting from removing respective subsidies. The level varies significantly between 0.1 USD/t in Poland and 3.4 USD/t for coal from the U.S. Powder River Basin (PRB). Depending on the producer this corresponds to less than 1% of production cost for Poland and South Africa, up to 34% for PRB coal.
The second policy examined in this chapter is a permanent moratorium on new coal mines, as suggested by President Tong of the Republic of Kiribati (Tong 2015) and supported by many scholars (see section 1.2.4). This policy could be implemented in various ways, e.g., by stopping to issue licenses for new mining projects and by not renewing those of inactive projects. To assess the consequences of such an intervention detailed information on existing mining operations is a crucial issue. There is a lack of publically available data, therefore I compile an own data set of reserves in operating mines based on publically available information. Based on this data, about one third of global reserves reported in international surveys (e.g., BGR 2015) are located in currently active mines. This share is largest in South Africa (69%) and smallest in the U.S. (8%).
Taking these two policies as scenarios, this chapter uses a comprehensive model of the world steam coal market COALMOD-World (see Chapter 2 for a detailed description of the model) to assess their effects on patterns of global steam coal trade, prices and CO2 emissions from coal consumption as
33 Namely, the scenario misses current developments in the U.S., China, and the EU. As an example, important regulations like the Clean Power Plan in the U.S. (EIA 2015b) are incorporated but not logically extrapolated to 2040. Moreover, the peak in coal consumption (NBSC 2015) and a moratorium on new coal power plants and mines in China are not accounted for (see The State Council of the People’s Republic of China (2016): “Coal Capacity Guideline Issued.” February 5. http://english.gov.cn/policies/latest_releases/2016/02/05/ content_281475284701738.htm., and Boren (2016): “China Stops Building New Coal-Fired Power Plants.” Energydesk. March 24. http://energydesk.greenpeace.org/2016/03/24/china-crackdown-new-coal-power-plants/). Likewise, the ban of coal from the energy mix in a number of European countries like in the UK is not included in the central scenario (cf. Rudd (2015): “Amber Rudd’s Speech on a New Direction for UK Energy Policy - Speeches - GOV.UK.” Gov.uk. November 18. https://www.gov.uk/government/speeches/amber-rudds-speech-on- a-new-direction-for-uk-energy-policy.).
61 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
well as their distribution effects. The two policies are assumed to be introduced in 2020. Although, generally the model works with perfect foresight, the policies are implemented in a way to ensure no anticipation effects. The subsidy removal policy leads to an insignificant reduction in CO2 emissions of, on average, 82 MtCO2 per year but still leaves a gap of 3.5 GtCO2 to be addressed by other measures to achieve emission reductions consistent with a 2°C target. Nevertheless, the policy generates considerable additional income for emerging countries (China 31.5 bn USD, India 8.1 bn USD, Indonesia 7.2 bn USD) in the period 2020 to 2040. This additional income can be used to finance additional measures to reduce CO2 emissions. Moreover, the policy generates additional revenue for infra-marginal producers that benefit from an average increase of coal prices by about 1% per year from 2020 to 2040, compared to the reference case. By contrast, a global moratorium on new mining projects could be a major contribution to closing the gap towards a coal consumption that is consistent with the 2° target. In fact, the “Mine Moratorium” scenario exceeds reductions implied by the WEO 450ppm scenario. The supply path in this scenario is, however, in line with McGlade and Ekins’ (2015) calculations on “unburnable” coal reserves. These are required to stay in the ground in order to achieve the 2°C target, without relying on CCTS.
The results of the two scenario analyses can be understood as a benchmark for the maximum ability of these policies to close the gap between the current consumption path and one that is consistent with the 2° target. The partial equilibrium setting of the underlying model does not specify the substitute that is used to compensate reduced steam coal consumption and therefore does not account for potential CO2 emissions from alternative sources. Also the model does not take into account welfare effects of recycling funds freed up by the removal of subsidies on coal production (see Chapter 2.7 for a discussion of model limitations).
The remainder of the chapter is organized as follows: the next section presents an overview of demand-side climate policies currently implemented and supply-side policies currently discussed. The subsequent section takes a closer look at coal producer subsidies and discusses findings from literature on their removal, and present own calculations on effects of subsidy removal. Section 3.4 discusses a moratorium on new coal mines as a potential supply side climate policy and details coal reserves in operating mines for the largest producers of steam coal. Furthermore, it gives a quantitative assessment of effects of a mine moratorium on the international steam coal market based on different specifications. Section 3.5 concludes.
3.2 Instruments of climate policy One common metric to categorize climate policies accounts for the side of the market for emission- intensive goods (in the scope of this chapter steam coal) that they address:34 those policies targeting the consumers are referred to as demand-side policies, while those addressing the production are referred to as supply-side policies (Kolstad et al. 2014, 364). Each policy has its specific advantages and disadvantages. Typical policy evaluation criteria assess the efficiency, the effectiveness, and the
34 This section heavily builds on earlier work published as Collins, and Mendelevitch (2015).
62 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
feasibility of a policy intervention (Perman et al. 2012). The Grantham Research Institute maintains a database of global climate legislation which details different policies that have been implemented (Grantham Research Institute 2015a).35
3.2.1 Demand-side policies
Demand-side policies for reducing CO2 emissions have received the most attention in the academic literature and have been most commonly introduced in practice. Carbon pricing instruments place an explicit price on emissions – either directly, as a carbon tax, or indirectly, through a cap-and-trade scheme (OECD 2013a). Such instruments have been implemented (or are scheduled to be implemented) in 39 countries, and at the jurisdictional level in a further three countries (Kossoy et al. 2015a, 22).
There are many other policy instruments which generate an implicit carbon price through regulatory intervention. Prominent examples are emissions performance standards, minimum flexibility requirements, renewable portfolio obligations (see Oei et al. (2014b) for a discussion of regulatory options to reduce CO2 emission in the power sector. Other demand-side policies include measures that promote energy efficiency and reduced energy consumption (as discussed in articles in Economics of Energy & Environmental Policy Symposium on “Energy Efficiency”: Gandhi et al. 2016; R. Hahn and Metcalfe 2016; Rosenow et al. 2016; Houde and Spurlock 2016).
In the absence of full participation in a global climate policy, demand-side policies are susceptible to carbon leakage: emissions-intensive activities shift to non-participating countries, such that emission reductions in the participating countries are partly offset by emissions increases in the non- participating countries (see e.g. Felder and Rutherford 1993; Sinn 2008). Richter (2015) provides an overview of empirical studies of the carbon leakage effect, which is undisputed in existence, but controversial in magnitude.
Moreover, a “green paradox” has also been theorized, where the expectation of future demand-side policies could induce resource producers to increase their present rates of extraction in order to maximize net present value (Sinn 2015). For coal, Haftendorn, Kemfert, and Holz (2012) suggest that in practice the green paradox may not be relevant, while Bauer et al. (2013) find a short term reduction of coal prices due to stringent climate policy. Gerlagh (2011) argues that the green paradox relies on oversimplified model assumptions with total depletion of the resource and high substitutability between energy fuels. Hoel (2012) adds that the paradox is only prevailing if policies target low cost suppliers while it is absent if it affects mainly high-cost suppliers of fossil fuel.
3.2.2 Supply-side policies Supply-side policies represent an alternative and more direct route to address negative effects of fossil fuel combustion. One important factor to consider when deciding between a demand-side and a supply-side policy is the ratio of demand vs. supply elasticity, as it drives the leakage risk for the
35 The following two sections are based on earlier work from Collins and Mendelevitch (2015).
63 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
respective policy. Lazarus, Erickson, and Tempest (2015) calculate this ratio for different fuels and regions based on various studies and find mixed evidence for supply-side and demand-side leakage risk for coal. Collier and Venables (2014) argue that for coal, supply-side policy may be less prone to leakage, and Hoel (2013) suggests the green paradox could be eliminated with a supply-side policy that targets high-cost coal deposits. Lazarus, Erickson and Tempest conclude that such climate policies are more likely to limit over-supply of fossil fuels and associated “carbon lock-in” effects.
One type of supply-side policy acts to directly remove coal reserves from production – whether to a partial extent (focusing on high-extraction-cost reserves for economic efficiency) (Harstad 2012), or to a further extreme, the progressive closure of the entire coal industry (Collier and Venables 2014). Another type of supply-side policy is a depletion tax (or alternatively, a depletion quota), which is analogous to the demand-side policy of a carbon tax (or for a depletion quota, a carbon budget). For instance, in Richter, Mendelevitch, and Jotzo (2015) propose a tax on the energy content of steam coal, levied by a coalition of major coal exporters. A supply-side policy for coal could also take the form of an export-licensing regime adopted by a coalition of major coal exporters, in analogy to the existing safeguards regime for uranium exports; based on the reasoning that the regulation of commodity exports on the basis of their harmful or unethical end use is a widely accepted principle, and should be extended to coal (A. Martin 2014). Lazarus, Erickson, and Tempest (2015) provide a comprehensive taxonomy of supply-side climate policies.
To date, there has been limited experience with the implementation of supply-side policies. The concept of preserving fossil fuel reserves has some precedent in the Yasuni-ITT Initiative, which was a proposal by the Ecuadorian government in 2007 to preserve oil reserves, but ultimately was not carried through (P. L. Martin 2014). A recent initiative that directly targets future coal supply is the “No New Coal Mines” campaign. It was started by the President of Kiribati who urged the leaders of the world to support this call for a moratorium on the opening of new and the expansion of existing mines (Tong 2015). This initiative is supported, inter alia, by the Obama administration (Warrick and Eilperin 2016) and by the Australia Institute (Denniss 2015b) which argues in favor of a global moratorium on new coal mines. Another supply-side policy which is broadly discussed at least since 1997 (cf. World Bank 1997) but only fragmentally implemented is a removal of fossil fuel subsidies. Both, the subsidy removal and the mine moratorium policy and their application to the steam coal market are discussed in detail in the two subsequent sections.
3.3 A production subsidy reform as a supply-side climate policy Influential country groups like the G20 (2009), APEC (2010), Friends of Fossil Fuel Subsidy Reform (GSI 2011), and UN Secretary General’s High-Level Panel on Global Sustainability (2012), have all committed to phasing out fossil fuel subsidies. Sustainable Development Goals, adopted in September 2015 by the UN (2015a) include a target focused on the rationalization of inefficient fossil fuel subsidies. To improve the understanding of the range and magnitude of fossil fuel subsidies in different countries, the Organization for Economic Co-operation and Development (OECD 2015b) conducted a comprehensive study. It counts almost 800 individual policies that support the production
64 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
or consumption of fossil fuels in OECD countries and six large partner economies (Brazil, the People’s Republic of China, India, Indonesia, the Russian Federation, and South Africa) with an overall value of 160-200 bn USD annually over the 2010-14 period. It estimates annual budgetary support and tax expenditure on coal subsidies to account for around 12 bn USD. A study by Ecofys which includes additional subsidy categories, found that coal subsidies in the EU-28 accounted for 10 bn EUR in 2012 (Ecofys 2014). Updating a global study by the International Monetary Fund (Clements et al. 2013), Coady et al. (2015) find fossil fuel subsidies accounting for 6.5% of global GDP (with 3.4%, or 2,530 bn USD originating from coal subsidies, with the major contribution of 2,506 bn USD due to global warming and local pollution externalities).
In developing economies, subsidy reforms are opposed by rent-seeking of incumbent stakeholders and divergence of interest between provincial and national governments (Dansie, Lanteigne, and Overland 2010). Often starting from a poor service level, governments are afraid to take unpopular decisions and induce social unrest. Citizens first need to be persuaded that the withdrawn support will be used in a welfare increasing way elsewhere.
Koplow (2015) provide a taxonomy of subsidies in energy industries. While they are commonly applied on both the demand and the supply side of fossil fuels, their removal may have very different effects and consequences depending on whether it affects producers or consumers. There is a large strain of literature analyzing the distributional incidence, induced emissions, and other distorting effects of demand-side fossil fuel subsidies (e.g. Arze del Granado, Coady, and Gillingham 2012; Dartanto 2013; Burniaux and Chateau 2014; Lin and Ouyang 2014; Schwanitz et al. 2014; Durand-Lasserve et al. 2015). Merrill et al. (2015) provide an overview of models examining the effect of fossil-fuel subsidy reforms on greenhouse gas emissions.
In this chapter, I want to concentrate on the implications of removing financial benefits granted to fossil fuel producers, and more specifically, coal producers. The removal of production subsidies for coal production can work as an effective supply-side climate policy. Such a policy comes with a double- dividend of removing heavy burdens from public budgets and reducing GHG emissions. Additionally, it can prevent carbon look-in by reducing capital-intensive investments from state-owned and international investors (Bast et al. 2015).
3.3.1 Definitions and data sources Article 1 of the WTO “Agreement on Subsidies and Countervailing Measures” (WTO 1994) defines subsidies as a financial contribution of a government or a public body that is directed towards a company or industry and involves i) direct transfer of funds, ii) foregone revenue (e.g., taxation below benchmark level) iii) provision of goods and services below market value, or iv) provision of funds or price support through indirect measures. This definition is non-judgmental on whether the measure is for some reason justified or efficient. Three major sources build on this definition and consistently estimate energy subsidies on a disaggregated level, but employ two contrary approaches for
65 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
assessing the respective subsidy level. The IEA’s36 definition centers on lowering costs or raising prices in a way that is beneficial for producers or consumers. The OECD37 (2015b) uses a similar definition but adds a reference to market levels. The IMF’s38 definition also distinguishes between pre- tax and post-tax subsidies, where the latter benchmarks to a price that also includes a “pigouvian” tax component correcting for externalities (Beaton et al. (2013) for a further discussion of different definitions of subsidies).
Figure 3.1: Illustration of different definitions of fossil fuel subsidies as a nested doll. Source: Adapted from Merrill (2014). Bárány and Grigonytė (2015) and Kojima and Koplow (2015) provide a comparison of the different methodologies to assess the magnitude of fossil fuel subsidies. The methodology used by IEA is the price-gap approach, which compares the end-user price to a reference price comprising free-on-board (FOB) costs, cost of shipping plus margins and taxes. The OECD method is based on the inventory approach, which concentrates on budgetary support and tax expenditures that entail merits for producers or consumers of fossil fuel, either relative to other activities or products, or in absolute terms. The IMF has adopted the price-gap approach in order to estimate pre-tax subsidies. Post-tax subsidies compare actual consumer prices with supply cost plus the efficient level of taxation which includes externalities and a fair consideration of margins. Due to these methodological differences IMF subsidy estimates are considerably higher than those published by IEA or OECD, as they also account for inefficient taxation of externalities (e.g. CO2, NOx emissions and local air pollution).
As neither the IMF data (IMF 2015) nor the IEA database (IEA 2016) distinguish between production and consumption subsidies, they cannot be used in this analysis. To the contrary, the method employed by the OECD is much more suitable, as it explicitly provides budgetary items that can be directly assigned to coal producers and their production costs. Where available, I use data from the OECD (2015a) and from ODI (2015c) that extends the effort undertaken by the OECD (2015a) and
36 See OECD/IEA (2016) for the exact definition. The IEA subsidy dataset is available at: http://www.worldenergyoutlook.org/media/weowebsite/2015/Subsidies20122014.xlsx. 37 See OECD (2015b) for the exact definition. The OECD subsidy dataset is available at: http://www.oecd.org/site/tadffss/. 38 See Coady et al. (2015) for the exact definition. The IMF subsidy dataset is available at: http://www.imf.org/external/np/fad/subsidies/data/codata.xlsx.
66 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
provide a detailed list by subsidy type, jurisdiction, fuel, and fuel chain stage. Import tariffs, like in the case of China (cf. Xue et al. 2015), constitute an indirect subsidy to domestic producers by lowering their exposure to competition on the world market. As the model framework used in this chapter does not account for this kind of market distortion, they are excluded from the analysis.
3.3.2 Findings from literature on coal production subsidies While many studies look into the effect of removing subsidies for all fossil fuels (see e.g. Schwanitz et al. 2014; Burniaux and Chateau 2014), there is only sparse literature on the effects of removing coal subsidies in particular. Anderson and McKibbin (2000) use the general equilibrium framework C- Cubed to assess the economic effects of removing production and consumption subsidies on coal. They examine two scenarios, one in which high income OECD countries remove domestic coal production subsidies and import restrictions at the same time. They find an average decrease of global
CO2 emissions of 5%. In the second scenario they additionally assume a removal of coal consumer subsidies and export taxes in Non-OECD countries and find an overall emission reduction of 8%. However, these strong results heavily rely on the authors’ “guess-estimates” of the subsidy levels, with subsidy removal increasing production costs by up to 250%.
Fulton et al. (2015) utilize a supply-demand partial equilibrium framework to derive aggregate supply and demand functions and assess the effect of adjusting the supply function by removing subsidies for coal in the U.S. Powder River Basin (PRB) as well as for Australian coal with a horizon from 2014 to 2035. Using a sensitivity analysis, they compute results for different demand elasticities and find that an increase of PRB supply costs by 4 USD leads to an annual emissions reduction of 21-55 MtCO2. The authors warn that unilateral removal of subsides again is prone to leakage effects.
3.3.3 Current subsidies on coal production in selected countries While there are various sources that report fossil fuel subsidy levels for different countries, the quality of available data differs substantially for observed countries. Comparing different sources, I have compiled a data set on steam coal production subsidies for eight major producers of steam coal, namely, USA, China, India, Australia, South Africa, Indonesia, Russia, and Poland (cf. Table 3.1). Identified subsidy levels range between 0.01 bn USD in Poland and 4.4 bn USD in China, for 2013- 2014. Per unit subsidies range between 0.1 USD/t for exported steam coal in Poland and 3.4 USD/t for steam coal produced in the Powder River Basin. A detailed description of the sources and of the calculation of subsidy levels for each of the analyzed countries can be found in Appendix A.7.
67 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
Total subsidies to Subsidy per unit of coal production in production and by Country 2013 [bn USD] region [USD/t] Comments USA 2.1 Powder River Basin 3.4 Forgone profits due to preferential tax Appalachia 1.1 treatment account for 50% others 1.0 China 4.4 Shanxi, Shaanxi, Inner Direct payments and investments, and Mongolia 1.3 the provision of services below market others 0.9 value account for 54%, and 39%, respectively. India 0.8 all 0.9 Investment by SOE Coal India Limited Australia 1.0 New South Wales 2.5 Lax treatment of rehabilitation liabilities Queensland 2.1 constitutes major subsidy others 1.8 South 0.04 transport to export Rail transport subsidy, below market Africa terminal 0.5 value sales to preferential consumers already disregarded in base case data Indonesia 0.9 all 1.8 Policies targeting to remove subsidies are not enforced Russia 0.07 0.4 Extreme divergence between sources on subsidy levels Poland 0.01 0.1 Free energy supply for mine workers Table 3.1: Total subsidy in 2013/2014 and subsidy per unit of production and by region for main coal producing countries. Source: Own compilation based on various sources. See country descriptions in the Appendix for details.
3.3.4 Quantitative assessment: production subsidy reform Quantitative results are obtained by employing the COALMOD-World model introduced in Chapter 2. The marginal cost intercept is adjusted according to the collected subsidy estimates reported in Table 3.3 to account for the removed subsidy. In the case of South Africa transportation costs between producer and exporter are adjusted, respectively.
The net effect of steam coal production subsidy removal on global CO2 emissions from steam coal is an emissions reduction of 2.5 GtCO2 (82 MtCO2 annually) for the model horizon until 2050. Roughly the same effect can be observed if Australia introduces a steam coal export tax of 18 USD/t (see Chapter 4) or if the U.S. unilaterally decides to introduce a moratorium on new coal mines on federal land (see Section 3.3.2). The effect can be considered insignificant, if compared to the required average annual reduction of 3.6 GtCO2, to close the gap between the WEO 2015 NPS and the 450ppm scenario (cf. Chapter 2). Table 3.2 reports results on producer, exporter, and consumer surplus, as well as total discounted level of removed subsidies.39 For the period 2020 to 2050, saved subsidies total 76 bn USD. While for the reformed countries producer surplus is reduced to a smaller extend than consumer surplus (24.7 bn USD and 34.3 bn USD, respectively), their net welfare effect is positive and totals 18 bn USD. The net effect for all examined countries is positive, except for India, due to its disadvantages role as a large net importer over the entire model horizon. As the subsidy
39 All monetary values are discounted to 2020, the year when the policy is assumed to be introduces. There is no anticipation of the policy in the preceding years, as the variables are fixed to “no policy” values for 2010 and 2015.
68 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
only affects export coal, consumers in South Africa are not affected by the policy. The policy induces an average price increase of 1% over the entire model horizon. South African, Russian, and Polish producers overall benefit from the policy as their cost increase is small relative to their competitors from Indonesia, Australia, and USA, therefore they exhibit a positive change in producer surplus.
In general, the removal of producer subsidies does not have a major impact on the steam coal market. Total saved subsidy volume accounts for 1.5% of total market volume over the model horizon. Though the net welfare effect is positive, it accounts for only 0.4% of total market volume over the same period.
Table 3.2: Effect of subsidy removal on producer, exporter, and consumer surplus. [bn USD 2020-2050, discounted CHN IND IDN USA AUS ZAF RUS POL to 2020] Total Total subsidy 31.5 8.1 7.2 23.1 4.5 0.8 0.8 0.1 76 Producer surplus -12.6 -2.3 -4.3 -6.2 -2.2 0.1 0.5 0.1 -24.7 Exporter surplus 2 - 0.4 -1.9 0 0.4 0 0 1 Consumer surplus -15 -6.3 -0.5 -7.1 -0.5 0.0 -0.8 -0.1 -34.3 Net welfare effect 5.9 -0.5 2.8 7.9 1.8 1.3 0.5 0.1 18
3.4 A moratorium on new coal-mines as a supply-side climate policy The “No New Coal Mines” initiative was started by the President of Kiribati who urged the leaders of the world to support this call for a moratorium on new and expansion of existing coal mines (Tong 2015). It is supported, inter alia, by Sir Nicolas Stern (Grantham Research Institute 2015b) and by the Australia Institute (Denniss 2015b), but also the U.S. and China have introduced a temporary moratorium on new coal mines (Warrick and Eilperin 2016; The State Council of the People’s Republic of China 2016). In addition to the usually quoted positive effects associated with reducing coal consumption, including environmental and health impacts, the proponents of a moratorium policy argue that it will also avoid stranded assets along the entire coal value chain and additionally reduce consumption through increased prices (Denniss 2015a; Finighan 2016). However, the policy comes with a caveat: Putting a moratorium on new coal mines gives a clear advantage to current incumbents and disadvantages new entrants (Denniss 2015a). In times of low coal prices and overcapacities on the market, the policy can also be understood as a classical industry support instrument. In the short- to medium-term a moratorium on new mines will stabilize prices, and thus generate revenue to current owners of resources and secure jobs and investments in current operations.40 At the same time, local economic benefits from new entrants and revenue from lease auctioning and royalties are foregone if a mine moratorium is implemented. Moreover, it potentially increases the carbon budget available to other fossil fuels, namely, to oil and gas.
40 E.g., Forsythe (2016) argues that current halt of coal-fired power plant construction and coal mines approval is rather due to economic reasons that to environmental concerns.
69 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
Literature that quantifies the effect of a moratorium on new coal mines is very sparse. Erickson and Lazarus (2016) examine the effect of phasing-out leases for fossil fuel extraction on government- owned land from which 40% of coal production currently originates in the U.S. For coal, their scenario assumes that currently issued licenses where production did not start are revoked and no new licenses are issued. They account for inter- and infra-fuel substitution and find that for coal such a policy could lead to emission reductions of 70 MtCO2/a, already corrected for a rise of 30 MtCO2/a from an increase in gas-fired electricity production. Finighan (2016) examines whether a global moratorium on coal mines would lead to a remaining coal budget that is consistent which the amount considered as “burnable” by McGlade and Ekins (2015). The latter uses a global energy systems model with a detailed representation of resources and reserves to assess the amount of fossil fuel that needs to remain in the ground to be in accordance with a 2°C target. Based on their assumptions on the costs and the availability of fossil reserves, 82-88% of coal reserves (and at least 96% of resources) must not be extracted. Finighan (2016) highlights that there is a lack of information on coal reserves in existing mines, which, however, would be required to test a “Mine moratorium” policy against the results obtained by McGlade and Ekins (2015). To overcome this lack, he employs two approaches: the first method uses a limited set of countries to estimate an average ratio of reserves to reserves in active mines and calculates 140 Gt of coal remaining in operating mines. The second method is based on the simplifying assumption that the lifespan of current mining operations is 20 years, and therefore current production levels could be maintained for 20 more years, if no new mines would be opened. Assuming an annual decrease in production of 5%, this method arrives at 126 Gt of coal remaining until 2050. Finighan (2016) finds that based on his estimates a mine moratorium would achieve a limitation of coal supply to volumes that are in line which the “coal budget” of 120 to 180 Gt calculated by McGlade and Ekins (2015) until 2050. However, the analysis has a number of drawbacks:
it relies on rule-of-thumb estimates of reserves in operating mines rather than a comprehensive data set, it does not allow to quantify the effects on market prices, trade patterns, and potential winners and losers of such a policy,
and it does not account for the heterogeneity of coal types and embedded specific CO2 emissions.
The following two subsections address some of the short-comings discussed above. Section 3.4.1 comprises an attempt to comprehensively collect data on coal reserves in operating mines for the major coal producing countries. Section 3.4.2 provides a quantitative assessment evaluating the effects on trade patterns, prices, CO2 emissions from coal, and welfare effects.
70 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
3.4.1 Remaining coal reserves in operating mines Due to individual assessment methods, prevailing complexity of measurement and measurement errors, as well as a political component, estimates of reserves and resources are hard to obtain and prone to substantial uncertainty. While there exists an international code for fossil fuel energy and mineral reserves and resources classification (UN 2013), it is not broadly used. Rather the code developed by the Joint Ore Reserve Committee (JORC 2012) is more and more commonly applied by companies, also outside its original Australasian scope. Based on various sources, BGR (2015) provides a comprehensive list of resource and reserves estimates for 81 countries. According to BGR (2015, 43), hard coal reserves totaled 699 Gt in 2014. A more in-depth, country-by-country analysis is available from the World Energy Council (2013) which reports a similar value of 691 Gt of proved recoverable reserves of anthracite, other bituminous and sub-bituminous coal by end of 2011. Thurber and Morse (2015b) and Osborne (2013), both provide a selected number of country case studies providing estimates of recoverable reserves and resources. The NGO “coalswarm”41 provides an incomplete list of mining operations in a limited number of countries. Commercial providers like “IntierraRMG”42 or “Mining Atlas”43 advertise to provide a comprehensive data set on operating mines globally, which, however, are not openly accessible. To my best knowledge, there is no comprehensive database that consistently reports remaining coal reserves in operating mines.
In the following, I present a comprehensive data set of coal reserves in operating mines44 on a country level. A detailed description of data origins and calculation methods can be found in Appendix A.8. Where available, mine level data was used based on publicly available data, inter alia, company reports, and ministry sources. For some countries, no such data could be acquired. Especially for China, due to a lack of available alternatives, I follow the methodology introduced in Finighan (2016). I apply an average quota of “reserves” to “reserves in operating mines” calculated based on available sources and apply it to reserves in China as reported in BGR (2015).
41 http://coalswarm.org/find-information/search-by-topic/coal-mines/. 42 http://www.snl.com/Sectors/metalsmining/Default.aspx. 43 https://mining-atlas.com/operation/php. 44 There is no clear-cut definition of coal reserves in operating mines. Where available, I rely on JORC code 111, Proved extractable reserves, reported for individual mine operations. I presume that extraction rights for these quantities are already acquired but I do not investigate the legal aspects in detail. Therefore, these figures might include coal reserves that are currently not developed but already considered a company asset. A further investigation of individual country mining and environmental law would be required to assess in how far such undeveloped reserves could also be retained in the ground without the need to adapt legislation and to cut into the legal rights of the individual companies.
71 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
Table 3.3: Estimates of resources and reserves from literature, and own estimates on reserves in operating mines.
BGR (2015) Estimated Estimated Reserves in COALMOD-World Reserves in 45 operating mines production node operating mines Resources Reserves [Gt] [Mt]
Australia 1536.7 62.1 19.8 P_AUS_QLD 681346 Colombia 9.9 4.9 3.2 P_COL 3221 China 5338.6 124.1 41.1 (85.2)47 P_CHN_SIS 2387448 P_CHN_Northeast 1836 P_CHN_HSA 8560 P_CHN_YG 6830 India 175 85.6 19.8 (48.4)49 P_IND_North 11607 P_IND_Orissa 6969 P_IND_West 1227 P_IND_South 500 Indonesia50 92.4 17.4 3.5 P_IDN 6122 Kazakhstan 123.1 25.6 2.2 P_KAZ 2200 Mongolia 39.9 1.2 n.a P_MNG 1170 Mozambique 21.8 1.8 n.a P_MOZ 212 Poland 162.7 16.2 0.8 P_POL 800 Russia 2658.3 69.6 17.7 P_RUS 17700 South Africa 203.7 9.9 6.8 P_ZAF 6800 USA 6457.7 222.6 17.6 P_USA_PRB 7050 P_USA_Rocky 755 P_USA_ILL 2821 P_USA_APP 4267 Ukraine 49 32 2.5 P_UKR 2600 Venezuela 6 0.7 n.a P_VEN 479 Vietnam 3.5 3.1 n.a P_VNM 150 Total in data 16878.3 676.8 135-207.7 Total in data base 128854 base World 17713.4 698.7 Share of 95% 97% world total Source: based on various sources as described for each country below.
45 Figures are adjusted and redistributed to coal basins covered by the COALMOD-World database. 46 Geosience Australia (2014) reports a spit of 7442/11547 between New South Wales and Queensland. I assume this ratio to remain constant. As COALMOD-World only covers international steam coal markets, numbers displayed apply the split between coking coal and steam coal using the current split in production figures with an average share of 59% for steam coal as reported by the Australian Government (2016) for the period 2009-2013. 47 The numbers are calculated based on the ratio of reserves reported by BGR (2015) to reserve in operating mines directly obtained from literature (for USA, Colombia, Poland, South Africa, Indonesia, and Australia). The number in brackets is based on the highest ratio obtained in South Africa (69%), while the standard assumption is the average ratio (33%). 48 Figures are distributed to the regional level based on the regional split employed in the COALMOD-World data base (see Appendix A, Table A.6). 49 The figure in brackets assumes that captive mine licenses are reissued while the standard assumption is that they are retired. 50 As the value of 3.5 Gt represents reserves as of end of 2015, to account for the model setting starting in 2010, the consumed amounts for 2010-2015 are added based on data from IEA (2015c; 2012b).
72 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
3.4.2 Quantitative assessment: Mine Moratorium Scenario Assuming an unanticipated reduction of available reserves to the levels reported in Section 3.4.1 in 2020 reduces total production by 42% (cf. Table 3.4). This corresponds to an emission reduction of on average 6.9 Gt per year for the period 2020-2050 or 5.2 Gt per year for 2010-2050. Annual CO2 emissions from coal in 2040 are 75% below the level observed in the reference case. The reserve constraint is binding for all steam coal producers, except Ukraine, Russia, and Australia Queensland producers, who can even expand their export compared to the reference case. This is due to the fact that these countries have low domestic consumption and have installed production at large deposits or just recently expanded production as in the case of Australia. Restricted reserves add a scarcity rent of on average 52.1 USD/t (production-weighted) to the price of coal. The policy leads to an average global price increase of 93% for the period 2020 to 2050. The global net welfare effect, disregarding any positive effect on climate change mitigation, is a 19% reduction in welfare, where a relative increase in producer surplus by 70% is outnumbered by a reduction in consumer surplus by 53%. The reduction in net welfare amounts to 18.4% of the steam coal market volume in the period 2020 to 2050. The highest reduction in consumer surplus can be observed in China, followed by India and USA. The policy comes with net benefits especially for Russia, Australia, and Colombia who profit from increased prices and reduced supply from competitors, especially from Indonesia. For South Africa there is a balance between positive price effects for exports and negative effects of a price increase on domestic consumption.
With tight reserve constraints, Chinese coal reserves are used up until 2040, while it increasingly relies on imports. Seaborne trade sees an even stronger concentration on China and India, while both domestic supply and imports to other countries is reduced by over 90%. Japan, Korea, Malaysia and Taiwan, are the only countries that have significant imports in 2040, besides China and India. USA consumption is reduced by on average 50%, with all reserves being used up by 2040. Similarly, South Africa uses up its reserves by 2040, Indonesia by 2035, and Poland by 2025. In total, international trade is reduced by 42% for the period 2020 to 2040.
Table 3.4: Cumulative production in the reference case and in the Mine Moratorium scenario in Gt. Cumulative production [Gt] Cumulative production [Gt] Mine Morat. Change Mine Morat. Change Country Base case Country Base case Scenario in % Scenario in % AUS 6.4 8.8 38 POL 1.9 0.8 -58 CHN 89.7 41.1 -54 RUS 6.6 10.6 61 COL 4.9 3.2 -35 UKR 1.5 2.1 40 IDN 13.0 6.1 -53 USA 32.3 14.9 -54 IND 30.1 20.3 -33 VEN 0.5 0.5 0 KAZ 3.5 2.2 -37 VNM 0.2 0.2 0 MNG 1.2 1.2 0 ZAF 12.0 6.8 -43 MOZ 0.2 0.2 0 Total 204.0 119.0 -42
73 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
3.4.2.1 Alternative specification: High estimate of reserves in operating mines This scenario assumes that available reserves in China and India are at the high estimate level reported in Table 3.3, which corresponds to an increase of 208% and 244%, respectively, compared to the values used in the “Mine Moratorium” scenario. In this scenario, results are in strong contrast to outcomes in the “Mine moratorium” scenario. Total production is only reduced by 18% compared to the reference case. This corresponds to an emission reduction of on average 2.1 Gt per year for the period 2020-2050 or 2.9 Gt per year for 2010-2050. Moreover, the scenario does not achieve a supply path that is consistent with the 2°C target as suggested by the WEO 450ppm scenario (cf. Figure 3.2). Due to the increased resource base, the reserve constraint is not binding for some regions in China and also for North India and Indian Orissa region.
As in the “Mine Moratorium” scenario, Australia, Russia and Ukraine do not deplete their reserves. Constraint reserves add on average 16.4 USD/t (production-weighted) to the price of coal, compared to 51.4 USD/t in the “Mine moratorium” scenario. The policy leads to an average global price increase of 33% for the period 2020 to 2050. Benefits for exporters, especially Russia and Australia, are 75% lower than in the “Mine Moratorium” scenario. The net welfare effect, disregarding any positive effect from climate change mitigation, is a 9% reduction in welfare, where a relative increase in producer surplus of 40% is outnumbered by a reduction in consumer surplus of 27%. This is less than half of the magnitude observed for the “Mine Moratorium” scenario. The reduction in net welfare amounts to 8.8% of the market volume of the steam coal in the period 2020 to 2050. With a reduction of 75% for the period 2020 to 2040, international trade is reduced twice as strong as in the “Mine Moratorium” scenario, due to more supply available on the domestic markets in China and India.
3.4.2.2 Alternative specification: McGlade and Ekins 2015 (M&E) scenario It is worth mentioning that, although both the WEO 450ppm scenario (2015a) and the calculations by McGlade and Ekins (2015), are based on an energy system that is consistent with the 2°C target, there is a strong divergence in the role that coal plays in these energy systems. While the former assumes that Carbon Capture, Transport and Storage (CCTS) is readily available, and has a share of 75% of total coal-fired electricity generation by 2040, the latter presents two specifications, including one without CCTS. There are two important caveats of the technology: First, CCTS increases coal required to produce the same amount of energy due to reduced efficiency (see Chapter 5 for more technical details on CCTS), and second, the technology is no available even at demonstration scale, yet (see section Chapter 3). Taking these issues into account, figures on future coal demand provided by the WEO 450ppm scenario likely need to be corrected downwards to be consistent with a 2°C target. Therefore, I also calculate coal supply patterns implied by coal reserves considered “burnable” by McGlade and Ekins (2015) in their specification without CCTS in this M&E scenario.
For the M&E scenario, I assume the introduction of a restriction of steam coal reserves as reported in Table A.13 in Appendix A.10, from 2020 onwards. For the period 2010-2020 there is no anticipation effect and the consumption is based on reserve data from the COALMOD-World dataset (cf. Table A.6 in Appendix A.1). Compared to the “Mine moratorium” scenario reserves are even more constrained,
74 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
but also the distribution of reserves is different. In the M&E scenario Poland has higher resource base, whereas reserves in Ukraine and Russia are substantially lower. Lower reserves are also assumed for South Africa, and Australia, while the resource base in Indonesia is almost at reference case levels. Finally, the split of reserves between China and India is different, with a larger share available to China.
Figure 3.2: Total steam coal consumption for different scenarios (in Mtpa). Due to similar total reserve base, results are in the same range as for the “Mine Moratorium” scenario. Total production is reduced by 47% (cf. Table A.12 in Appendix A.10) compared to the reference case, which corresponds to an emission reduction of on average 7.8 Gt per year for the period 2020-2050 or 5.8 Gt per year for 2010-2050. The reserve constraint is binding for all steam coal producers, except for Poland. The constraint adds on average 56.4 USD/t (production-weighted) to the price of coal. The policy leads to an average price increase of 102% for the period 2020 to 2050. The net welfare effect, disregarding any positive effect from climate change mitigation, is a 21% reduction in welfare, where a relative increase in producer surplus of 61% is outnumbered by a reduction in consumer surplus of 23% and exporter surplus of 51%. The reduction in net welfare amounts to 19.4% of the market volume of steam coal markets in the period 2020 to 2050. More reserves available in Indonesia are outnumbered by reductions in Russia and Ukraine. Therefore, international trade cannot compensate for additionally tightened reserves in India. Consequently, production levels in 2040 are even below those in the “Mine Moratorium” scenario. International trade is reduced by 53% compared to the reference case for the period 2020 to 2040.
75 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
3.5 Conclusions Reducing coal consumption is one of the core means to achieve the 2°C target. Observing frustration on the outcomes achieved by demand-side climate policies in the past two decades, supply-side policies represent an alternative approach which can complement demand-side climate change mitigation efforts. In this chapter I investigate the effect of two supply-side climate policies on consumption, prices, and patterns of trade on the international steam coal market and domestic coal markets.
The first policy follows the suggestions of the G20 (2009) and other influential groups and examines the effects of removing subsidies for steam coal production. The policy comes with a double-dividend, by first removing heavy burdens from public budgets and, second, reducing greenhouse gas emissions. I find subsidy levels ranging from 0.1 USD/t in Poland to 3.4 USD/t for U.S. coal from the Powder River Basin. While I find a positive welfare effect of removing these subsidies of in total 18 bn
USD for the period 2020 to 2050, the effect on CO2 emissions from coal can be considered insignificant for a global policy. The calculated average annual reduction of 82 MtCO2/a only makes up for a small fraction of the 3.6 GtCO2/a required to be consistent with the 2°C target.
Still, the removal of production subsidies for fossil fuels can work as an effective supply-side climate policy. However, such a reform should not be considered as an isolated measure but as part of an integrated climate policy package. On the contrary, if accompanying policies aimed at internalizing fossil fuel externalities, are not implemented across fuels, a pure subsidy reform can even lead to an increase in domestic coal consumption, like investigated for Indonesia by ADB (2015).
As the definition of subsidies is non-specific on whether a subsidy is for some reason justified or suited to correct for market failure, the figures used in this chapter also include measures such as compensation payments for mines shut down in the Chinese “Coal Phase-Out Plan” (cf. Appendix A.7). These payments may be well justified as they reduce output in the long-term and provide a transition period to mitigate negative effects on local small scale firms. To provide an integrated cost- benefit analysis for each of the policy interventions interpreted as subsidies is beyond the scope of this chapter. The figures presented should rather be interpreted as first attempt to consistently assess the economic and environmental effect of removing coal production subsidies on global coal consumption and trade patterns on the global steam coal market.
The set of subsidies included in this analysis does not include the costs induced by not accounting for externalities caused by the production and consumption of coal which can be understood as a social subsidy. This is common practice by IMF (2015). Estimates of these “social costs of carbon” are difficult to obtain, but are increasingly incorporated into policy and other impact assessment studies. Depending on the discount rate and timing of the emissions EPA reports “social cost of carbon” between 11USD/tCO2 and 95 USD/tCO2 (EPA 2015b). Including such additional costs would have a significant effect on coal consumption, but also on trade.
The second policy that is investigated in this chapter is a moratorium on new and expansion of existing mines as suggested by Tong (2015), President of the Republic of Kiribati, but also other scholars. The
76 Chapter 3: Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?
policy again comes with a double-dividend: First, it achieves emission reductions by conserving reserves, and second, compensating current resource owners through increased scarcity rents and therefore market prices. Due to a lack of consistent data on reserves in operating mines, I compile my own data set based on publicly available data. Total reserves in these mines are estimated at 137.3- 210 Gt, depending on assumed reserves in India and China.
While the high estimate of remaining reserves fails to achieve a consumption pattern in line with the WEO 450ppm scenario, the “Mine Moratorium” scenario, assuming the lower estimates exceed required reductions. The supply path in this scenario is, however, in line with McGlade and Ekins’ (2015) calculations on “unburnable” reserves coal. These are required to stay in the ground in order to achieve the 2°C target, without relying on CCTS. In the “Mine Moratorium” scenario, prices increase by on average 93%, while total production is reduced by 42%. Not taking into account the positive effects of reduced emissions of CO2 and other local pollutants as well other local externalities, the positive effect to producers is outnumbered by a decrease in consumer welfare, leading to a net welfare reduction of 19%.
While, on the long run, a permanent mine moratorium can be a significant contribution to climate change mitigation, the policy comes with a serious caveat: In the short- to medium-term, it is particularly beneficial for current incumbents and disadvantages new entrants. Therefore, such a policy should not be considered to be introduced in isolation. Otherwise, there is a risk that it will be deemed a temporal industry support policy that protects current incumbents without any long-term effect on reducing CO2 emissions.
Both examined policies are very much suited to be applied in a broader scope, covering not only coal but eventually all fossil fuels. The effect of the timing of the introduction and a potential expansion of the policies across fossil fuels should be further investigated. It is likely to govern in how far inter-fuel competition can be used to temporally align incentive and create favorable conditions to introduce such policies.
77 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
Chapter 4 COAL TAXES AS SUPPLY-SIDE CLIMATE POLICY: A RATIONALE FOR MAJOR EXPORTERS?
4.1 Introduction: A climate policy with a double dividend Coal is both the fossil fuel with the highest carbon intensity per unit of energy, and an energy carrier that is globally abundant (Rogner et al. 2012).51 Even though coal use is in decline in most developed countries, over the past decade, it has regained momentum and a “renaissance of coal” ensued (Steckel, Edenhofer, and Jakob 2015; Schernikau 2010). After the current pause, global coal use is expected to continue growing on account of rising demand in developing countries, notably India (IEA 2015a; OECD/IEA 2015).
This is in stark contrast to the climate policy imperative: if climate change is to be kept within tolerable limits, most of the proven global coal reserves need to remain in the ground (Meinshausen et al. 2009; McGlade and Ekins 2015). The vast majority of climate policies are designed to reduce the demand for fossil fuels in general and coal in particular. However, the price-dampening effect of such policies put fossil fuel exporters at a disadvantage, which reduces their willingness to join global efforts to mitigate climate change. By contrast, supply-side climate policies - that reduce fossil fuel use through supply constraints - could potentially leave energy exporters better off through improved terms-of-trade. Moreover, they can be attractive to implementing governments as they raise fiscal revenue which could, among other uses, be deployed to help with the transition to low-carbon energy systems.
In this chapter, we focus on the case of coal taxes, investigating both the incentives for implementation, and the impacts of withholding supply. Specifically, we consider hypothetical taxes on the export or production of steam coal52 that are levied by Australia, the world's second largest steam 53 coal exporter, or alternatively levied by a coalition of major exporters. CO2
51 This chapter is submitted to Journal of the Association of Environmental and Resource Economists. A previous version was published as DIW Berlin Discussion Paper No. 1471, 04/2015. It is joint work with Philipp M. Richter and Frank Jotzo. Roman Mendelevitch and Philipp M. Richter jointly had the lead role in the model development, the implementation in GAMS, and the analysis of results. Philipp M. Richter had the lead role in the writing of the manuscript. 52 Steam coal includes all hard coal that is not coking coal (used for steel production), as well as sub-bituminous brown coal (IEA 2013b). It is mainly used for electricity generation and has by far the largest share in global extraction across the different types of coal. 53 A debate about constraints on expanding coal exports in the interest of climate change mitigation is nascent in Australia, but to date has not been underpinned by quantitative analyses. The implementation of an Australian export tax for climate reasons has been suggested for instance by Peter Christoff http://theconversation.com/why- 78 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
We thereby contribute to the literature in various ways: First, in contrast to the existing, mainly theoretical, literature, we numerically analyse in great detail one particular supply-side climate policy.
Specifically, we investigate short-term and long-term effects of coal taxes on CO2 emissions, tax revenues, and shifts in the global patterns of consumption, production, and trade of steam coal. Second, we focus on coal. While coal is the most relevant fossil fuel for climate change mitigation, it has gained surprisingly little attention in the literature so far.54 We seek to fill this gap. Third, we set up a model that represents the value chain of the steam coal sector in great detail. To this end, we build on, and further develop, COALMOD-World (see Chapter 2), which replicates global patterns of coal supply, demand and international trade. Our model features endogenous investments in production and transportation capacities in a multi-period framework and represents the substitution relation between imports and domestic production of steam coal. Hence, in response to coal taxes, short-run adjustments (e.g., import substitution effects) and long-run reactions (e.g., capacity expansions) of competing exporters and importing countries are endogenously determined in the model. We embed this model as the lower level in a two-level game, where at the upper level policy makers maximize tax revenues by levying coal taxes on exports or production. Fourth, we follow recent developments in operations research and develop an algorithm to numerically solve this Mathematical Program with Equilibrium Constraints (MPEC). As solution techniques applied to large-scale models are still in a development stage (cf. Siddiqui 2011; Gabriel and Leuthold 2010), we combine and test different methods in order to robustly solve the outlined problem. The modelling approach developed in this chapter can be applied to markets for other commodities also.
This chapter contributes to different strands of literature. First, we add to the literature on supply-side climate policy that was initiated by Sinn (2008) who highlights the intertemporal trade-off of resource owners in the face of demand-side climate policies:55 If climate policy became increasingly strict over time, resource owners would rationally react with early extraction leading to accelerating global warming - a green paradox. For this reason, climate policy should tackle the supply side in order to slow down the extraction path of fossil fuels and to incentivise the conservation of carbon in the ground.
In a seminal paper, Harstad (2012) analytically investigates a compensation scheme for resource-rich countries through the introduction of a market for extraction rights. Committing to conserve coal
australia-must-stop-exporting-coal-9698, and Brett Parris http://theconversation.com/expanding-coal-exports-is- bad-news-for-australia-and-the-world-17937. A tax review proposed a resource rent tax including coal (Commonwealth of Australia 2010), which was legislated in 2012 but repealed in 2014. Relatedly, Martin (2014) suggests a coal-export safeguard regime encompassing the major steam coal exporters to ensure that trade only occurs with partners that offset related CO2 emissions. 54 Recently, and in line with the perception of a 'renaissance of coal', the literature on international steam coal markets gained some attention: In the tradition of Kolstad and Abbey (1984), two numerical models of the international steam coal market have been developed using equilibrium modelling techniques, one model by Trüby, and Paulus (2012), for the other one see Chapter 2. It is the later that we develop further. These models have been applied to analyze the impact of steam coal transportation in China on the global market (Paulus and Trüby 2011), and to investigate different climate policy scenarios (Haftendorn, Holz, and Hirschhausen 2012). We complement this work by analyzing coal taxes as supply-side climate policy. 55 See Lazarus, Erickson, and Tempest (2015) as well as Chapter 0 for up-to-date overviews of the research on supply-side climate policy.
79 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
deposits in situ, a climate coalition can cost-efficiently reduce emissions without provoking carbon leakage; resource owners in turn generate revenues by selling extraction rights.56 Hoel (2013) finds that preventing the extraction of the most expensive reserves reduces overall emissions and does not provoke a green paradox. Hagem and Storrøsten (2016) assess how supply-side policies implemented by a “coalition of the willing” influences extraction decisions of other countries. Collier and Venables (2014) finally extend the, Harstad (2012) idea by including details on enforcing the coalition through moral pressure and a ring-fence cap-and-trade system, where oil resource owners compensate owners of coal deposits to leave the coal in the ground.
In contrast to proposed compensation mechanisms for reduced extraction, we focus on the complementarity of rent capturing and climate change mitigation: Levying coal taxes, generates, on the one hand, revenues against the background of improved terms-of-trade - a motive for trade policy well-known from the literature. On the other hand, the implementation of an export tax represents a climate policy instrument as it reduces global coal consumption.
Our contribution is thereby complementary to Fæhn et al. (2017), who derive the optimal mix of demand and supply-side climate policies for the resource-rich Norway. While focusing on the crude oil market, they find that the largest contribution to meet CO2 reduction targets should be made by withholding oil extraction. In contrast to Fæhn et al. (2017) we rely on a detailed representation of the supply side (of steam coal) including capacity constraints and trade costs. Supply curves are hence endogenously derived and not determined by assumptions on long-term supply price elasticities.
While we represent the international steam coal market as being competitive, we assume that governments of major exporting countries may well affect their terms-of-trade through trade policy, similar to Kolstad and Abbey (1984) and Kolstad and Wolak (1985). This chapter is, hence, related to the literature on the interaction between climate policy and strategic behaviour. Based on the work of Tahvonen (1995), Liski and Tahvonen (2004) construct a game of two strategically acting groups of players: A supply-side cartel and a coalition of importing countries. They show that the optimal carbon tax of the importing countries includes trade policy elements in order to extract rents from the supply- side cartel.57 Böhringer et al. (2014) find that the extent of carbon leakage depends on the degree of market power on the supply side of fossil fuels. In particular, carbon leakage is sensitive to the assumed behaviour of OPEC. Furthermore, Persson et al. (2007)and Johansson et al. (Johansson et al. 2009) argue that resource owners like OPEC may benefit from a global carbon price if the marginal price is set by other energy carriers which are more carbon-intensive and, hence, more affected by
56 It is the conservation motive and purchase of extraction rights by a climate coalition that makes the total supply inelastic at the socially optimal level. Hence, there is no need for additional policy instruments. In contrast, Bohm (1993) and Hoel (1994) – that Harstad (2012) is building on - show that depending on demand and supply price elasticities an optimal mix of demand and supply-side policies may similarly help to avoid carbon leakage. In addition to being less prone to carbon leakage, supply-side climate policies may additionally have distributional advantages (cf. Asheim 2013). 57 Kalkuhl and Brecha (2013) highlight the importance of the distribution of climate rents, which are created by the climate policy-induced scarcity of remaining CO2 emissions. In this context, Eisenack et al. (2012) find that - depending on the allocation rule---a global carbon cap may indeed leave resource owners better off compared to a business-as-usual.
80 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
climate policies. In contrast to these papers that focus on demand-side climate policies in the context of supply-side cartels, we analyse the case for coal taxes levied by a coalition of export countries or a ‘grand coalition’ of all coal producing countries.
Finally, this chapter adds to the literature on numerical applications of MPECs that can be found in a wide range of disciplines and research questions. In contrast to analyses of market power with similar players on both the upper and lower level (cf. Siddiqui and Gabriel 2013; Trüby 2013; Gabriel and Leuthold 2010), our framework allows for the analysis of economic policy decisions given the reaction of market participants.58 To the best of our knowledge, there has been no study that numerically investigates the climate and distributional effects of a supply-side climate policy implementation (at the upper level) on an international fossil fuel market (at the lower level).
Our main results suggest that for Australia a positive and substantial export tax of about 7 USD/tCO2 would maximize the net present value (NPV) of tax revenues. The total (discounted) tax revenue could reach more than 16bn USD in the next 20 years, associated with reduced consumer prices in Australia, while the average world price for steam would increase. However, over the simulation period until 2035 we find a strong rebound effect with on average more than 73% of reduced Australian exports being compensated by increased production in importing countries (most notably India and China), and additional exports from competitors on the international market. This high rebound effect attenuates a large impact on global coal use and CO2 emissions.
Though also prone to leakage, a coalition of the four major exporters of steam coal, Indonesia, Australia, Colombia, and South Africa, could have a significant effect on coal markets by cooperatively setting an export tax. At the revenue-maximizing level of about 10 USD/tCO2, CO2 emissions from steam coal use would be reduced by on average 190 Mt annually. We test the sensitivity of the results to the members of the coalition by adding the USA. We find a substantial increase in tax revenues and
CO2 emission reductions, while the US share in additional revenue is small.
By contrast, we find production-based coal taxes, which affect domestic supply and exports alike, less prone to rebound effects. Across all scenarios, revenue-maximising rates for coal production taxes are consistently higher than for export taxes, associated with larger tax revenues and lower global CO2 emissions than export-only taxes. Levying production taxes further avoids distortions between exported and domestically used coal in exporting countries.
To put the role of coal taxes in perspective, we further investigate what level of coal taxes is necessary to reduce global coal use consistent with the 2°C target. This can only be achieved by a globally levied tax on production. However, even a moderate tax rate of 10 USD/tCO2 could significantly reduce emissions from coal use by 0.07 to 1.9 GtCO2 annually---depending on the regional coverage of the tax.
58 Other applications of optimal policy decisions in the context of two-level games include Bard et al. (2000) on the optimal incentive to encourage the production of biomass, Labbé et al. (1998) on the optimal toll setting for a road network, and Scaparra and Church (2008) on the cost-efficient way to protect a service system from being disrupted by saboteurs or terrorists.
81 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
The remainder of the chapter is organised as follows. Section 4.2 presents the upper level setup and the solution algorithm developed for numerical application. Section 4.3 provides a description of the analysed scenarios. Results are discussed in Section 4.4. Section 4.5 concludes.
4.2 Model description and specification Consider a two-level game with an optimal policy problem at the upper, and a multi-period equilibrium model of the international steam coal market at the lower level. We assume that the coal tax is introduced by an economic decision maker g who can anticipate the equilibrium reactions of all market participants, namely producers fF , exporters eE and final consumers cC of steam coal. In turn, these economic players take the policy decision parametrically.59 We follow Kolstad and Abbey (1984) and model the choice of the tax rate by the policy maker to be based on tax revenue maximization.
In the following model description we focus on the mathematical formulation that applies to the case of coal taxes on exports. The corresponding mathematical formulation for coal taxes on total production can be found in Appendix A.3, where we also provide a list of all endogenous variables, and the complete model formulated as equilibrium problem by Karush-Kuhn-Tucker (KKT) conditions.
4.2.1 Upper level: Policy maker as Stackelberg-leader
E At the upper level, policy maker g can levy a tax a on steam coal exports in periods aA in order
E to maximize the NPV of tax revenues. While the policy maker can decide on the initial tax rate 0
60 starting in period aa , the path is predetermined by an annual growth rate of r :
E E a a a 0 (1 r ) (4.1)
Accordingly, the policy maker's optimization problem is given by
a 1 E max aEXP aec (4.2) E 0 aec 1 rg
where EXPaec subsumes total exports of all exporters e being located in decision maker ’s territory, and directed towards consumption nodes c , which are outside ’s territory. It is the impact of the tax on total exports that can be anticipated by the policy maker. Periodic revenues are
59 This Stackelberg-leader-follower relation is a common assumption in the literature (cf. Eisenack, Edenhofer, and Kalkuhl 2012) and requires the existence of a credible commitment of the respective policy maker (cf. Brander and Spencer 1985). 60 By varying this growth rate, the analysis is not restricted to a constant tax rate but rather allows for an endogenous starting level that gradually increases over time. We provide sensitivity analyses of the revenue- maximizing tax rate regarding the growth rate of the tax in Appendix A.5. Note that the alternative to implement the tax rate as a free variable for each period could create time inconsistencies.
82 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
discounted at rate rg . Note that we model the export tax based on the energy content of exported volumes; it is hence proportional to a carbon tax.
4.2.2 Lower level and data set specifications The upper level described above is added to the COALMOD-World framework detailed in Chapter 2. The mathematical formulation can be found in Appendix A.1. The latter constitutes the lower level of
E the two-level problem, where the upper level variable a is incorporated into the respective exporters’ and export-oriented producers’ problem, as detailed in the mathematical formulation.
In general the data set introduced in Chapter 2 is used to perform the numerical analysis of the two- level problem in this chapter. However, there are two major deviations: First, the model horizon is limited to 2035; and second, Latin American consumers are not included into the analysis. Both is due to a recent update of the model data set for Chapter 2 which is not incorporated in this chapter, but rather it is based on an slightly older COALMOD-World version published in Holz et al. (2015).
4.2.3 Solution algorithm In mathematical terms, the two-level problem described by Equation 4.2 and Equations (A.9)-(A.18), (A.23), (A.25)-(A.30) is an MPEC. Solving MPECs numerically is a demanding task due the combination of an optimization problem at the upper, and an equilibrium problem at the lower level which pose fundamentally different restrictions on potential solutions.61
The most readily available option to solve MPECs using the software GAMS is to rely on the commercial solver NLPEC. The upper level, in our case represented by Equation 4.2 together with a KKT reformulation of the lower level problem constitute the MPEC that is input to the NLPEC solver (see Appendix A.1.2-A.1.4 for the full set of KKT conditions given by Equations (A.9)-(A.18), (A.23), (A.25)-(A.30)). This solver works computationally fast and has been applied in the literature (cf. Trüby 2013; Huppmann and Holz 2012). However, this solution method is opaque and does not necessarily provide the user with the globally optimal solution. In order to circumvent drawbacks of the NLPEC
e method, we vary the upper and lower bounds of decision variable 0 in a multitude of runs to obtain different local optima as candidate solutions. We pick the candidate with the highest objective value as our global optimum. As a robustness check we further take a one level formulation of our problem and compute equilibrium results for a grid of different initial tax rates. The shape of the resulting distribution of tax rates suggests that the optimum obtained from the NLPEC runs is indeed global.
An established alternative solution technique for MPECs is to reformulate the lower level problem into a Mixed-Integer Problem (MIP) by using disjunctive constraints (see Fortuny-Amat and McCarl 1981; for the original formulation, and Gabriel and Leuthold 2010, for a recent application). We develop an algorithm that combines these two methods and apply it to smaller datasets. See Appendix A.4 for
61 See Luo et al. (1996) who provide an overview of different solution techniques for MPECs that has been thoroughly updated by Siddiqui (2011).
83 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
further details. However, due to extensive computational burdens, we rely on the NLPEC solver for the large-scale numerical application as that is detailed in the following sections.
4.3 Scenario definitions The Base Case of our model is constructed in line with the New Policies Scenario (NPS) of the World Energy Outlook 2012 (IEA 2012d). The NPS is a scenario of moderate climate policy, assuming the implementation of current energy and climate policy proposals. While CO2 emissions decline in some regions (e.g., by 16% between 2010 and 2035 in OECD countries), global emissions are on an increasing path. Accordingly, global steam coal consumption is projected to rise through 2035. China and India jointly share more than 70% of global coal consumption throughout. While we base our reference coal consumption levels on IEA (2012b), the patterns of production and international trade flows are endogenously determined in the model representing profit-maximising firm behaviour.62
We construct two main coal tax scenarios and analyse differences to our Base Case. Both scenarios rely on the imposition of coal taxes on exports with different countries that levy the tax:
The first scenario Tax AUS focuses on a unilateral export tax on steam coal levied by Australia.63 Australia has a large share in international trade of steam coal (about 16% in 2012 according to the IEA (2013b) ) and its policies are perceptible on the international coal market. The second scenario Tax Coalition analyses the situation of a co-ordinately set export tax by a coalition of major exporting countries, namely Australia, Indonesia, Colombia and South Africa that together had a share in international traded steam coal of 72% in 2012 (cf. IEA 2013b).
We vary these two main scenarios along different dimensions: First, we test the sensitivity of the Tax Coalition scenario to the members of the coalition by adding the USA. Second, we compare the levy of export taxes to the case of taxes levied on the entire production of coal. Third, we calculate revenue- maximizing export and production taxes for grand coalitions of all exporters64 and all producers, respectively. Finally, and in order to assess the tax level required to reduce global coal use consistent with a 2°C target we derive a boundary scenario based on the WEO 450 ppm scenario (IEA 2012d). This boundary case serves as benchmark to evaluate induced reductions in global coal use in our different scenarios.
For all scenarios, we assume a discount rate of 10% for producers and exporters and a government discount rate on tax revenues at rg 0.05 , with a moderate annual tax growth rate rt 0.025 as default values. In order to test for sensitivity, we vary these parameters (see Appendix A.5). For all scenarios, sensitivity runs and robustness checks, we assume the introduction of the scenario-specific coal tax in the model period 2015 and rule out anticipation by any model agent: all variables in the
62 Reference consumption levels, reference prices and demand price elasticities are used to derive the linearly approximated inverse demand curves in Equation (A.30). 63 While we distinguish the two Australian coal producing regions New South Wales and Queensland, the same tax rate applies to both regions. 64 The exporters grand coalition consists of Australia, Colombia, Mozambique, Indonesia, Poland, Russia, South Africa, Ukraine, the USA, and Venezuela.
84 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
previous period 2010 are fixed at Base Case levels. We hence avoid any inconsistencies and do not allow for adjusted infrastructure expansions in 2010 in anticipation of the tax rate.
4.4 Discussion of results In this section we present our numerical results as follows. In Section 4.4.1 we discuss in depths both export tax scenarios. These results are contrasted against the levy of production taxes in Section 4.4.2. Results on the boundary case and the grand coalition scenarios are presented as part of the general comparison of results in Section 4.4.3. Finally, we qualify our results and discuss limitations in Section 4.4.4.
4.4.1 Export taxes on coal Levying an export tax on coal can lead to four partial effects on patterns of coal consumption and production. First, coal extraction and exports from the tax-implementing country are reduced, potentially accompanied by an increase in world coal prices (the terms-of-trade effect). Second, supply on the domestic market gets more attractive compared to the alternative of exporting; hence, consumption in the tax-setting country increases. Third, production in all other countries may increase- --a rebound effect: Export competitors compensate for the lower international supply, while net importers rely on domestic production to a larger extent. Note that this is a specific form of carbon leakage different to the well-established energy market channel (cf. Burniaux and Oliveira Martins 2012). It is not the increased consumption of fossil fuels in non-regulating countries encouraged by lower prices that increases emissions, but, on the contrary, an increase in the supply in non-regulating resource-rich countries incentivised by rising prices. Fourth and finally, according to increases in world 65 coal prices, global consumption of coal - and thus global CO2 emissions from coal use - is reduced.
The following discussion of scenario results focuses on these tax-induced changes in the patterns of consumption, production and trade relative to the Base Case with a particular emphasis on CO2 emissions and carbon leakage.66
4.4.1.1 Scenario Tax AUS: A unilateral Australian export tax on coal According to the optimization problem laid out Equation (A.32) the revenue-maximising Australian 67 export tax on coal starts at a level of 6.7 USD/tCO2. This is a significant level that is equivalent to a tax of about 18 USD per tonne of Australian steam coal. By comparison, average prices for coal exported from Australia were around 58 USD/t in August 2015 (IEA 2015a). In 2015, 225 Mt of steam coal were exported from Australia,68 so the tax take would have been around four billion USD, or around one percent of the total tax take by the Australian federal government. In the modelling
65 Note that we do not model substitution effects into other fuels which would partly offset the CO2 emissions effects from coal. For a discussion on the limitations of our modelling approach see Section 4.4.4. 66 Guiding Base Case results can be found in Table A.6 in the Appendix. 67 Calculations in the model are in terms of energy units (GJ or PJ). The conversion to a CO2-content based measure assumes an emission intensity of 0.0983tCO2/GJ. 68 Department of Industry, Innovation and Science http://www.industry.gov.au/resource/Mining/ AustralianMineralCommodities/Pages/Coal.aspx, accessed on June, 30 2016.
85 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
scenario, given the assumed government discount rate of 5%, the NPV of tax revenues until 2035 is around 16 bn USD, 69with only small changes in annual tax revenues.70
Figure 4.1: AUS – Export tax: Australian production, consumption and exports in Base Case and in Tax AUS (in Mt). Figure 4.1 shows how Australian consumption, production and exports are affected by this tax. While in the Base Case, Australian exports increase over time by almost 50% between 2010 and 2030, in Tax AUS exports decrease from their 2010 level until 2020, before remaining at about the same level throughout the model horizon. In every period exports are significantly lower than in the Base Case, approaching only 50% in 2030.
Since exporting gets relatively more expensive due to the additional costs incurred by the export tax, supply to domestic consumers is encouraged in Tax AUS. Here, consumption levels are higher by up to 14% above the Base Case value in 2030. This is in accordance with Australian consumer prices being well below the Base Case level.
As the effect on Australian consumption is small compared to the export-reducing effect, production levels are reduced significantly. The gap between Base Case production levels and those under Tax AUS is more than 60 Mt in 2030, or about one third of Base Case levels. Moreover, the reaction to an Australian export tax is visible in the production levels of all other countries (RoW). Figure 4.2 depicts changes in production levels for all relevant countries.
69 For the calculation of cumulative and average values we rely on a linear interpolation between the model periods. Furthermore all cumulative values are calculated for the period 2015 to 2035. 70 For a sensitivity analysis of the NPV of tax revenues with respect to the discount rate see Appendix A.5.
86 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
Figure 4.2: AUS – Export tax: Changes in supply to the international market (left figure) and to domestic markets (right figure) relative to the Base Case (in Mt). When the tax is introduced, exporting competitors are not able to rapidly expand their supply as most of them are already running at capacity. The exception is some additional supply from Russia which is redirected from its domestic market, and more significantly from the USA. At the same time large consumers, namely China and India, are not able to increase their domestic production on short notice. It is only by 2020, that the capacity can be expanded, and domestic production gains a comparative advantage over high-price imports made possible by sufficient domestic transport capacity. By then, China and India combined produce an additional 30 Mt of steam coal annually.
Figure 4.3: AUS – Export tax: Changes in patterns of global consumption relative to the Base Case (in Mt). Similarly, after 2020, international export competitors of Australia increase their supply. After 2025, Indonesia and Russia together compensate for around 25% of reduced Australian exports. The USA,
87 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
Colombia and South Africa cannot competitively increase their exports significantly, which can partly be explained by the long distances to the main destinations of Australian coal in Asia. In the longer run around 35-45 Mt of reduced Australian annual exports remain uncompensated by other net exporting countries. Hence, world steam coal trade is significantly reduced relative to the Base Case. By contrast, annual global production is reduced only by a small amount, on average by 12 Mt, down from 24 Mt in the starting year, highlighting a pronounced rebound effect
Figure 4.4: AUS – Export tax: Decomposed impact of the Australian export tax relative to the Base Case, in in Mt, and change in weighted CIF prices in percentage (right axis). Country-level consumption patterns change for different reasons (see Figure 4.3). First, importing countries generally reduce consumption as a result of an Australian export tax. In particular, consumption is significantly lower in China and India despite the increase in domestic production. Consumption is lower in exporting countries such as in Russia or Indonesia due to a shift in supply from the domestic to the export market. Second, and as discussed above, consumption in Australia is higher due to shifts from exporting to supplying domestic consumers. In line with consumption, global 71 CO2 emissions from coal are lower than in the Base Case: Annual emissions from coal use decrease 72 by up to 63 MtCO2 (in 2015); on average by 36 MtCO2 per year until 2035.
73 At the revenue-maximising tax level, we find average abatement tax revenue of 22 USD/tCO2. That means---abstracting from second order effects---that the Australian government could gain 22 USD for every ton of CO2 from coal usage that is reduced. By contrast, if reductions in coal use were brought
71 Note that consumption-based emissions increase in Australia but global emissions decrease in line with global coal consumption. 72 Put into perspective this corresponds to 15% of GHG emissions from the Australian energy sector in 2013. See http://ageis.climatechange.gov.au/, accessed on May 20, 2015. 73 We define average abatement tax revenue as NPV of tax revenue divided by total reduction in CO2. This is in contrast to the standard term of abatement costs and highlights the complementarity between emission reductions and tax revenue generation.
88 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
about by measures in importing countries---that is, through demand side rather than supply side policies---Australia would see only lower prices and lower volumes without additional tax revenue.
Figure 4.4 summarises the impact of an Australian export tax decomposed into exports and production for the domestic market and differentiated between Australia and all other countries. The large shift in global steam coal production and a significant rebound effect (on average of 73%) is observable.74 Additionally, Figure 4.4 depicts the effect of the tax on average consumer prices for steam coal worldwide. With an increase of 4% in 2015, the price increase is the strongest, when the market cannot completely adjust to the tax shock. By contrast, there is hardly any price change in the periods after 2015.
Figure 4.5: AUS – Export tax: NPV of tax revenues, in bn USD, as well as change in global CO2 emissions from coal use, in Gt (right axis), as a function of the initial export tax rate, in USD/tCO2. In summary, the unilateral levy of an Australian export tax significantly changes global patterns of consumption, production and exporting - but mainly in the short-run. It only has a small impact on total global coal use and related CO2 emissions, even at levels above the revenue-maximising tax rate (cf. Figure 4.5)
4.4.1.2 Scenario Tax Coalition: A joint export tax on coal by major exporting countries A coordinated export tax by the four major exporting countries Indonesia, Australia, South Africa and Colombia leads to a stronger terms-of-trade effect and higher tax revenues compared to any unilateral
74 In the Appendix, we depict global steam coal trade flows in the year 2030 for the Base Case and both export tax scenarios, Tax AUS and Tax Coalition, in their default specification.
89 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
75 76 policy action. The common revenue-maximising tax level is 10.1 USD/tCO2. This tax rate is significantly higher than the unilaterally introduced Australian export tax.
The NPV of joint tax revenues over the next 20 years is about 125 bn USD; the Australian share---in the absence of any redistribution scheme---is about 16 bn USD, which is similar to the revenues in the unilateral Tax AUS case. depicts the NPV of tax revenues as a function of the initial export tax rate and differentiated by coalition members. Depending on the tax level, Indonesia generates 45-65% of the annual tax revenue in this coalition.
At the revenue-maximising tax level, global coal exports are reduced by 20% while the coalition restrains about 40% of its exports. Globally, steam coal exports are reduced since other exporting countries cannot compensate significantly for the supply restrictions of the major four exporters.
Figure 4.6: Coalition – Tax revenues of the coalition’s members, in bn USD, and change in cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of the initial common production tax rate, in USD/tCO2. As in Tax AUS the rebound effect, however, is mainly driven by a pronounced increase in production destined for domestic markets---in particular in China. Nevertheless, and as expected, global consumption is reduced to a larger extent than in the unilateral Australian tax case. On average, global CO2 emissions are reduced by 194 MtCO2/a, compared to annual reductions of 36 MtCO2 in the
75 As a sensitivity check we calculate the revenue-maximising tax level for a coalition that additionally includes the USA and analyse resulting effects on prices and emissions. The discussion of these results is deferred to the Appendix A.6. 76 This corresponds to a charge of approximately 22-26~USD per tonne of exported steam coal, depending on the carbon content of the respective coal that differs across producers.
90 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
unilateral case. Over the next 20 years more than 3.9 GtCO2 could be emitted less than in the Base Case.
Figure 4.7 summarises the volume effect of the coalition's export tax, and the impact on average consumer prices for coal. The relative price increase is most pronounced in the year, when the tax is introduced. Between 2015 and 2020 it drops from 8% to 1% but then again steadily increases to 3% in 2030. The price trend is inversely related to the restrained supply by the coalition members.
Figure 4.7: Coalition – Export tax: Decomposed impact of an export tax jointly set by the coalition of major exporters relative to the Base Case, in Mt, and change in weighted CIF prices in percentages (right axis). Figure 4.8 shows that, for any given tax level, the Tax Coalition case always has higher average revenues per tonne of CO2 avoided (a market power effect) and higher cumulative emission reductions than the Tax AUS case. Note that for the revenue-maximising tax rate the level of average additional revenues is at about 22 USD/tCO2 (at cumulative CO2 emissions reduction of 0.7 GtCO2) for the Tax AUS case, while it is 32 USD/tCO2 (at cumulative CO2 emissions reduction of 3.9 GtCO2) for the Tax Coalition case.77
77 We summarise important results for all scenarios in Table A.10 in Appendix A.2 and in Section 4.4.3.
91 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
Figure 4.8: Tax AUS and Tax Coalition: Comparison of average revenues per ton of CO2 abated, in USD/tCO2, and reduction in cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of the initial value of the export tax, in USD/tCO2.
4.4.2 Production taxes on coal
The analysed coal taxes on exports lead to a net reduction of global CO2 emissions. However, under an export tax coal consumption is shifted to domestic consumers in tax-setting countries, causing increases in domestic consumption and economic inefficiencies. Export taxes are also criticized for their trade-diverting effects and might see WTO challenges. For this reason, we alternatively model taxes levied on the entire production of coal in the respective tax-setting country.78
4.4.2.1 Scenario “Tax AUS”: Export vs. production tax Figure 4.9 shows that there exist two peaks in the NPV of tax revenue with respect to a production tax levied by Australia. The first one lies at an initial value of the coal tax of 8.8 USD/tCO2. This is the same order of magnitude as for the case of an export tax, while emission reductions are more pronounced as both Australian exports and consumption levels decline. Here, tax revenue is generated from payments of both domestic and foreign consumers. From this first peak, a further increase in the tax rate then reduces export supply and ultimately leads to the situation where there are no coal exports and domestic consumers carry the entire tax burden.79
78 Investigating a tax on produced rather than exported volumes requires changes in the mathematical formulation of the model. These changes are described in detail in Appendix A.2.2. 79 Note that these results are sensitive to our assumptions on demand price elasticities and the linearly approximation of the inverse demand function around a reference price-quantity equilibrium. Moreover, since Australia is included as exporting country only, there is no alternative supply to the domestic consumers who may want to switch to importing steam coal.
92 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
Figure 4.9: AUS – Production Tax: Australian tax revenues, in bn USD, and change in cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of the initial production tax rate, in USD/tCO2. This second peak, or global optimum, is reached at a much higher initial tax level of almost 40
USD/tCO2. Up to this maximum, the increase in the tax rate outweighs the decline in quantities. Tax revenues are substantially higher; emissions reduced to a larger extent; production in 2030, in turn, is reduced to a tenth compared to the Base Case. However, it is only domestic consumers who carry the tax burden.
4.4.2.2 Scenario “Tax Coalition”: Export vs. production tax A tax levied on the entire production in Colombia, South Africa, Australia, and Indonesia, is a major intervention in the cost structure of the world steam coal market affecting 17% of global production and 72% of global exports (2010 values from the model dataset, based on (IEA 2012b)). Similar to the unilateral case, we find tax revenue maximisation to occur at a production tax rate that exceeds the coalition's export tax, in this case by 20%, reaching 12.2 USD/tCO2.
Figure 4.10 depicts the decomposition of tax revenues across members and shows how global emissions decline in the initial production tax level. For the same reasons as in the unilateral case, tax revenue as a function of the initial tax level peaks twice. Contrary to the unilateral tax scenario, however, here the first peak yields the highest revenues.
In contrast to an export tax, the coalition's supply to its domestic market is substantially reduced. Overall, both production and export levels are reduced by almost 50% in 2030 relative to the Base
Case. Cumulative CO2 emissions until 2035 are reduced by almost 8.5 Gt at the revenue maximizing tax rate. In this scenario South Africa, with its high and increasing domestic coal demand, has a dominate role in the coalition, with almost 50% of the coalition's revenue (compared to 15% in the
93 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
export tax scenario). The NPV of the coalition's total revenue is 266 bn USD which is twice the amount of the export tax scenario. Consumers in the coalition countries experience an average price increase of 10%, compared to a global average increase of 5% relative to the Base Case.
Figure 4.10: Coalition – Production Tax: Tax revenues of the coalition’s members, in bn USD, and change in cumulative global CO2 emissions from coal use, in Gt (right axis), as a function of the initial common production tax rate, in USD/tCO2.
4.4.3 Comparison of results Table A.10 in Appendix A.2 summarises key results on tax revenue, production, trade and prices, as well as emissions across the different scenarios presented in the previous sections. For the unilateral and small coalition scenarios we find revenue-maximising tax levels within a range between 6.7 and
12.2 USD/tCO2, while on average emission could be reduced between 37 to 423 MtCO2 annually. The production tax scenarios consistently show higher revenue-maximising tax levels and CO2 emission reductions, while carbon leakage rates are consistently lower compared to their export tax counterparts. This result can be explained with market distorting effects of export taxes, which make domestic consumption relatively cheaper. Moreover, average price increases are higher for production taxes, while, by contrast, we find similar results for the average abatement tax revenue for production and export taxes. Across the three Tax Coalition scenarios, the production tax case yields higher average emission reductions paired with less pronounced price effects. This suggests that the market's reaction to a tax set by a large coalition is similar to the impact of a smaller coalition which applies the tax to both exports and domestic sales.
Additionally, Table A.10 in Appendix A.2 presents results of the boundary scenario and grand coalition scenarios. The analysis reveals that a consumption pattern that is consistent with the WEO 450ppm scenario (average CO2 emission reductions of 4.0 GtCO2/a) is neither achieved by an export tax (1.2
94 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
GtCO2/a), nor by a production tax levied by all coal exporting countries only (2.6 GtCO2/a), at the tax revenue maximizing rates. It is only when a global regime of coal taxes is implemented, that the level is clearly exceeded (7.6 GtCO2/a). When comparing the market effects of the different scenarios, the supply-side policies led to strong price increases (16%-112%), while in the case of global demand-side driven reduction in steam coal consumption, prices decrease by on average 17%.80 The tax levels of
25.9-42.6 USD/tCO2 is well in the range of 1-130USD/tCO2 for carbon tax rate observed globally, but well above the 10 USD/tCO2 of 85% of all carbon taxes in place (Kossoy et al. 2015b). At the same time, the higher value of revenue-maximising tax levels is just in the range of the price required to drive coal-to-gas switching in Europe is around 40EUR/tCO2 estimated by Gray (2015).
Finally, we compare results for all scenarios using a common tax rate on coal of 10 USD/tCO2 (see Table 2). Doing so we can properly highlight the difference in the size of the coalition (ranging from one country only to grand coalitions) and between coal taxes on exports and the entire production. For instance, if a coalition of major coal exporters set an export tax, then reductions in CO2 emissions from coal use are five times larger than if Australia as a major exporter acted alone. Moreover, if all coal producing countries implemented this moderate tax rate of 10 USD/ tCO2, global emissions could be reduced by almost 2 GtCO2, annually - a significant contribution to climate change mitigation.
4.4.4 Qualification of results We have applied a comprehensive tool to analyse the global steam coal market at a disaggregated level. Nevertheless, we consider it necessary to put our approach into perspectives.
Firstly, at the lower level of our two-level modelling exercise, we rely on a partial equilibrium model of the international steam coal market. Substitution with other fossil fuels is only indirectly taken into account through inverse demand functions. A relative price increase in coal - e.g. through the introduction of coal taxes - would partly ramp up the consumption of natural gas and crude oil. The effects on CO2 emission reductions from lower coal consumption would hence be partly compensated by more emissions from other sources. Nevertheless, since coal products are the most carbon- intensive fossil fuels, global emissions would likely be reduced. In this sense, our analysis gives the upper bound of emission reductions.
Secondly, in all scenarios---independent of whether the tax is based on exports or production---the path of the tax rate is exogenously given, starting from an endogenous initial level, though. This modeller's choice reduces complexity and avoids time inconsistencies of jumps in the tax rate. In order to attenuate this shortcoming, we analyse a variety of possible developments of the tax rate and provide some additional results in the Appendix A.5.
Thirdly, we focus on tax rates that maximize fiscal revenue. This is distinct from welfare maximisation; indeed we make no attempt to quantify welfare effects. There are two reasons for this choice: On the one hand, this chapter highlights the complementarity between CO2 emission reductions and tax
80 Note that in all scenarios, prices are derived from an endogenous cost escalation mechanism, which takes into account depletion of cheap resources, and continuous investment requirements due to depreciation of assets.
95 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
revenue. On the other hand, existing exact solution techniques for large-scale MPECs require linear objective function at the upper level which limits our choice to focus on tax revenue. Nevertheless, we additionally provide results for tax levels other than the revenue-maximising rate---both graphically and in Table A.10.
Finally, in our setting only one country, or alternatively one group of countries, can act as Stackelberg- leader. We consequently neglect possible reactions of other economic decision makers beyond market driven reactions. These may include retaliation of importing or other exporting countries through their own taxes, tariffs or trade restrictions. Mathematically, representing more than one player at the upper level constitutes an Equilibrium Problem with Equilibrium Constraints (EPEC). Solving EPECs is a fundamentally different task than solving MPECs, while existing solution methods cannot be applied to large-scale models.
4.5 Conclusions This chapter investigates supply-side climate policy through coal taxes. We analyse to what extent large coal exporting countries can help to achieve global climate change mitigation, while at the same time benefitting from tax revenues against the background of improved terms-of-trade. To this end, we construct a two-level game with a policy optimisation problem at the upper level, and an equilibrium model of the international steam coal market at the lower level. We analyse numerically how restrictions of coal supply impact the international coal market through reactions of competing exporters and importing countries, and to what extent these can reduce global coal use and resulting
CO2 emissions.
While Australia may unilaterally generate additional tax revenues, we find a strong rebound effect with more than 73% of the reductions in Australian exports being compensated by increased production from competing exporters and domestic production in India and China. The net emissions reduction is significant in short term, and almost non-existent in medium to long term as other producers fill the gap. In contrast, our results suggest that a coalition of the major exporters Indonesia, Australia,
Colombia, and South Africa is necessary to significantly lower global CO2 emissions. Accordingly, we find that global CO2 emissions could be reduced by almost 4 Gt over the next 20 years in case of a co- ordinately levied coal tax on export at 10 USD/tCO2.
International trade law and the possibility of retaliatory trade action by importing countries speak against the introduction of such export taxes and the formation of a cartel of major exporters. A production-based tax may not face the same difficulties, and our results indicate that it could bring higher tax revenues and greater emission reductions. For instance, if the same group of major exporters imposed a tax on production rather than emissions, global emissions could be reduced by more than twice. Alternatively, if all coal producing countries implemented a moderate tax rate of 10
USD/ tCO2 on production, an annual emissions reduction of almost 2 GtCO2 could be achieved.
In a nutshell, our results suggest that coal taxes---at levels that are beneficial for exporters - can reduce global CO2 emissions. But even a grand coalition of all steam coal exporting countries would not optimally choose to levy a tax that achieves global emission reductions consistent with the 2°C
96 Chapter 4: Coal Taxes as Supply-Side Climate Policy: A Rationale for Major Exporters?
target. Nevertheless, supply constraints by fossil fuel exporters may hold promise as part of a broader climate change mitigation strategy. Crucially, they can leave the owners of fossil fuel reserves better off, in contrast to the conventional policy approach of tackling the demand side.
This chapter sets aside questions of political feasibility and of the distribution of rents among countries agreeing on joint coal taxes and between different interest groups within countries, such as companies and taxpayers. Similarly, neither do we investigate the stability of the analysed coal tax coalition, nor do we consider the various options for revenue recycling, such as potentially welfare-enhancing tax reform or redistribution. Moreover, our partial equilibrium framework does not allow for the evaluation of economy-wide effects on production, consumption and investment that would be brought about by coal taxes. Further analyses could usefully explore the induced shift to other fuels and its impact on
CO2 emissions in order to gain a more holistic view on the role that coal taxes can play as supply-side climate policy. We leave all this for future research.
97
Part B CARBON CAPTURE, TRANSPORT, AND STORAGE
98 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Chapter 5 MODELING A CARBON CAPTURE, TRANSPORT, AND STORAGE INFRASTRUCTURE FOR EUROPE
5.1 Introduction: the impact of the carbon capture, transport, and storage technology The ongoing Carbon Capture, Transport, and Storage (CCTS) discussion originates from multiple perspectives: On the one hand, longer-term energy system models insist on the need of CCTS to achieve ambitious decarbonization scenarios (IEA 2009).81 On the other hand, progress in advancing the technology on the ground has been modest thus far (Herold, Hirschhausen, and Ruester 2010; Hirschhausen, Herold, and Oei 2012c). The IEA underlines in its “Energy Technology Perspectives 2012” study that its importance with on overall 20% contribution to achieving emission reduction goals and an 40% cost increase in absence of the technology (IEA 2012a). At the same time they acknowledge the real danger that the ambitious development plans for CCTS demonstration in Europe will remain unfulfilled. Among other concerns, the institutional question about regulatory and environmental issues with storage could substantially hinder the deployment. In December 2012 the European Commission decided not to consider any CCTS project in the first round of the NER300 funding program, but supporting 23 renewable energy projects with €1.2 bn, instead. The lack of financial guarantees from project partners and member states as well as insufficiently advanced project status highlighted the uncertain future of CCTS in Europe (EC 2012).
To date, the discussion has centered on the role of CCTS in the power sector (Tavoni and Zwaan 2011), yet the technology also holds promise for the iron and steel, cement as well as refining sectors where chemical processes emit large amounts of CO2. Switching to renewable sources and/or increasing process efficiency will result in partial emission reductions in the medium term, e.g., 35% in the iron and steel sector, 35% in cement and 20% in clinker production (Öko-Institut 2012). Low- carbon substitutes to the conventional production of these raw materials, such as magnesium cement or the electrolytic production of iron, may become available in the future. However, the extent to which they could be applied on a large scale as well as prove economically viable is unknown. Thus the
81 This chapter is published in the Journal of Environmental Modeling and Assessment 05/2014; December 2014, Volume 19, Issue 6, pp 515-531 (Oei, Herold, and Mendelevitch 2014). Previous versions were also published in Zeitschrift für Energiewirtschaft Volume 35, Number 4, p. 263-273, 2011 (Oei, Herold, and Tissen 2011) and as DIW Berlin Discussion Paper No. 1052, 09/2010 (Mendelevitch et al. 2010; Oei et al. 2010). Joint work with Pao- Yu Oei and Johannes Herold. Roman Mendelevitch and Pao-Yu Oei jointly developed the model, and its implementation in GAMS. Andreas Tissen was also involved in developing a first draft of the model. The writing of the manuscript was executed jointly.
99 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
CCTS technology remains the only short-and midterm CO2 mitigation option for these sectors. At the same time an application in these sectors will lead to lower capture costs than in the energy sector due to the higher CO2 concentration in the flue gas (Öko-Institut 2012; Ho, Allinson, and Wiley 2011).
Despite this fact industrial partners have made only little effort to bring forward CCTS projects. Most industrial companies also lack the financial possibilities to invest into a demonstration unit including transport and storage of CO2. One major argument against putting pressure on industrial facilities in Europe, is the fear of losing international competitiveness when facing higher production costs due to CCTS. This apprehension was, for example, present in the design of the allocation scheme for EU- ETS emission allowances. The pure grandfathering approach did not put any pressure on the emission efficiency of existing facilities (as widely criticized e.g. in IETA 2012) and thus free allowances were used instead of pushing for CCTS. The only two large-scale industry CCTS demonstration projects in Europe, ULCOS Florange (a steel making plant in Lorraine, France), and Green Hydrogen (a hydrogen plant in Rotterdam, Netherlands) initially applied for NER300 funding, but then, in 2012, withdrew their application (MIT 2016). It is worth noting that those industry CCTS projects that are currently operating face favorable and very site-specific conditions. Either CO2 capture is disproportionally inexpensive due the specific process (e.g. Ethanol Production, in Decatur,
Illinous, USA), or the CO2 has to be captured regardless in order to market the product (e.g. natural gas with a too high CO2-concentration as in e.g. Sleipner field, Norway), or additional revenue from
CO2 enhanced oil recovery changes the economics of the CCTS project (e.g. Weyburn Project in Saskatchewan, Canada). Von Hirschhausen et al. (2012b) analyzed the discrepancy between the hopes put into the technology and its state of development (see Chapter 5). In addition to the points mentioned above, they found that there was a lack of technological focus on cheap capturing technologies. Also, too optimistic expectations on cost reductions and learning curves, as well as the fact that the costs and complexity related to regulatory issues of CO2 transport as well as regulatory and technological issues of CO2 storage were neglected. Moreover, persisting negotiations and complicated environmental assessments for CO2 storage fueled by “not in my backyard” (NIMBY) concerns hindered the implementation of planned demonstration projects. Against this background the question arises of what contribution the CCTS technology can realistically make toward European CO2 emission reduction.
We apply the CCTS-Mod Model to analyze the potential development of a CCTS infrastructure in
Europe. In particular, we investigate the nature of the CO2 transport infrastructure that would emerge in Northwest Europe, i.e. in Germany and its neighboring states. Several scenarios, differing by the estimate of geological storage available, the availability of onshore storage, and the expected CO2 certificate price in 2050, are run. We find that under certain extreme assumptions, such as a relatively high CO2 price, and very optimistic CO2 storage availability, a large-scale CCTS roll-out might indeed be expected. However, in a more realistic scenario, including lower storage availability and public resistance to onshore storage, a large-scale roll-out is not going to happen. In all scenarios, CCTS deployment is highest in CO2 intensive non-energy industries, where emissions cannot be avoided by fuel switching or alternative production processes.
100 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
The next section 5.2 provides an overview of existing literature and models, both theoretical and applied, e.g. to North America or Europe. Section 5.3 specifies our own model, called CCTS-Mod and its data. We then apply CCTS-Mod to analyze the potential development of a CCTS infrastructure in Europe under certain scenarios in section 5.4. Section 5.5 summarizes the findings and provides conclusions.
5.2 Modeling CO2-infrastructure Recent literature points out that the real bottlenecks to CCTS deployment are transport and storage infrastructure (Herold, Rüster, and Hirschhausen 2010). Against this background, only a few simplified
CCTS models actually address the pipeline transport of large volumes of CO2. The Global Energy Technology Strategy Program (GTSP) modeled the adoption of a CCTS system within three fossil fuel-intensive electricity generation regions of the U.S. The results show that CCTS implementation depends more on CO2 injection rates and total reservoir capacity than on the number of potential consumers who would use the CO2 for enhanced oil recovery (CO2-EOR) (Dooley et al. 2006).
McPherson et al. (2009) and Kobos et al. (2007) introduced the "String of Pearls" concept to evaluate and demonstrate the means for achieving an 18% reduction in carbon intensity by 2012 in Texas using CCTS. Their dynamic simulation model connects each CO2 source to the nearest sink and automatically routes pipelines to the next neighboring sink, thus creating a trunkline connection for all of the sinks. While the model can determine an optimal straight-line pipeline network, it is not possible to group flows from several sources to one sink. Fritze et al. (2009) developed a least-cost path model connecting each source with the nearest existing CO2 sink. The chapter examines a hypothetical case of main trunk lines constructed by the U.S. Federal Government and its influence on the total costs. However, no economies of scale are implemented for construction, as the costs of building the public trunk lines are greater than the potential costs of private enterprises. Nevertheless public trunk lines allow greater network flexibility and redundancy which can lead to cost savings in times of emergency and when storage capacity needs to be balanced.
Middleton et al. (2007) designed the first version of the scalable infrastructure model SimCCS based on mixed integer linear programming (MILP). With its coupled geospatial engineering-economic optimization modeling approach, SimCCS minimizes the costs of a CCTS network capturing a given amount of CO2. An updated version by Middleton and Bielicki (2009), comprising of 37 CO2 sources and 14 storage reservoirs in California, simultaneously optimizes the model according to the amount of
CO2 to be captured from each source; the siting and construction of pipelines by size; and the amount of CO2 to be stored in each sink. The decisions are endogenous, but the total amount of CO2 to be stored is exogenous. Economies of scale are implemented via possible pipeline diameters in four-inch steps, each with its own cost function. Kuby et al. (2011) extend a smaller version of the model that employs twelve sources and five sinks in California with a market price of CO2 as well as a benefit when used in CO2-EOR. This model minimizes the costs of CCTS, but only examines one period.
Their findings of a CO2 price sensitivity analysis indicate that infrastructure deployment is not always sensitive to the price of CO2. Kazmierczak et al. (2008) and Neele et al. (2009) develop an algorithm to create a low-cost network and a decision support system to evaluate the economical and technical
101 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
feasibility of storage. A realistic estimate of the economic feasibility of a potential CCTS project is possible, but there is no detailed planning at the project level.
In summary, only a few models include economies of scale in the form of possible trunk lines, but they operate on a static level or are based on an exogenously set amount of CO2 to be stored. Therefore the models exclude the option of buying CO2 certificates instead of investing in CCTS infrastructure. We introduce a scalable mixed integer, multi-period, welfare-optimizing CCTS network model, hereafter CCTS-Mod. The model incorporates endogenous decisions on carbon capture, pipeline, and storage investments as well as capture, flow and injection quantities based on exogenous costs, a
CO2 certificate price path, a comprehensive set of emissions point sources from European power and industry sectors as well as on- and offshore storage sites in depleted hydrocarbon fields and saline aquifers. Our model runs in five-year periods beginning in 2005 and ending in 2050. Capacity extensions can be used in the period after construction for all types of investments in the model. Sources and sinks are linked to nodes according to their geographical position and pipelines are constructed between neighboring nodes. To ensure a better resolution no aggregation of sources/sinks at a node takes place. The distance between two neighboring nodes can be chosen flexibly, making CCTS-Mod scalable and thus allowing different degrees of resolution. Economies of scale are implemented by discrete pipeline diameters with respective capacities and costs.
5.2.1 Mathematical representation of CCTS-Mod
Figure 5.1 illustrates the decision path of CCTS-Mod based on the CO2 disposal chain when using the
CCTS technology. Each producer of CO2 must decide whether to release it into the atmosphere or store it via CCTS. The decision is based on the price for CO2 certificates and the investment required for the capture unit, the pipeline and the storage facilities, and the variable costs of using the CCTS infrastructure.
Figure 5.1: Decision tree in the CO2 disposal chain of the CCTS-Mod Source: Own depiction. We apply a stylized institutional setting to a vertically integrated CCTS chain. A single omniscient and rational decision-maker has perfect foresight and makes all investment and operational decisions.82 Under these simplifying assumptions we run the model using a single cost minimization.
82 The model tends to overestimate the potential for CCTS. Considering the large number of different players and technologies, the uncertainties regarding CO2 prices, learning rates, legal issues, permitting, certification of 102 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
We define the objective function to be minimized as follows:
1 (푦푒푎푟푎−푦푒푎푟푠푡푎푟푡) min ∑ [( ) 푥푃푎,푖푛푣_푥푃푎, 1 + 푟 푧 ,푓 ,푖푛푣 , 푎 푃푎 푖푗푎 푓푖푗푑푎 푝푙푎푛푖푗푎,푦푆푎,푖푛푣_푦푆푎
⋅ (∑[(푐_푐푐푠푃푎 + (1 − 푐푎푝푡_푟푎푡푒) ⋅ 푐푒푟푡푎) ⋅ 푥푃푎 + 푐_푖푛푣_푥푃 ⋅ 푖푛푣_푥푃푎 + 푐푒푟푡푎 ⋅ 푧푃푎] 푃
+ ∑ ∑ [퐸푖푗 ⋅ (푐_푓 ⋅ 푓푖푗푎 + ∑(푐_푖푛푣_푓푑 ⋅ 푖푛푣_푓푖푗푎푑) + 푐_푝푙푎푛 ⋅ 푝푙푎푛푖푗푎)] 푖 푗 푑
+ ∑[푐_푖푛푣_푦푆푎 ⋅ 푖푛푣_푦푆푎])] 푆
(5.1)
With:
푥푃푎, 푖푛푣_푥푃푎, 푧푃푎, 푓푗푖푎, 푦푆푎, 푖푛푣_푦푆푎 ≥ 0 (5.2)
푖푛푣_푓푖푗푑푎 ∈ 푁0 (5.3)
푝푙푎푛푖푗푎 ∈ [0; 1] (5.4)
The first term of the objective function 1 is the discount factor, where r is the interest rate, yeara is the starting year of period a and start is the starting year of the model. From here, we can split the objective function 1 into three parts representing the three steps of the CCTS chain. For the first step the decision variables are the dimensioning of the capture system inv_xPa and the level of CO2 emission that are cycled through the capturing system (xPa · capt_rate represents the amount of CO2 actually captured by the facility). An individual variable is declared for each emitter P in each period a.
The parameter capt_rate represents the maximal possible percentage of captured CO2, thus certificates still have to be purchased at the price of certa for the remaining fraction. It is kept constant at 0.9 for all scenarios.
The second part represents the transportation step. The decision variables are: fija declares the CO2
flow from node i to j in period a; inv_fijda denotes the number of pipelines to be built between node i and j with the diameter d in period a; planija is a binary variable (see Equation 4) and has the value one if a pipeline route between node i and j is planned and licensed in period a, and zero otherwise.
storage capacity, and further policy measures would increase the total costs. Real costs are therefore expected to be higher and come along with a lower deployment of CCTS in the future.
103 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Routing of pipelines is a central aspect of our study; we implement a detailed process of pipeline building by introducing the planning variable and, thus, separate the planning and development costs from the rest of the capital costs. Additional pipelines on already licensed routes do not face licensing or planning costs. The desired effect is that new pipelines are routed along existing lines as observed in reality.
The third part represents storage. The decision variables are: ySa, which is the quantity stored in storage facility S in period a, and inv_ySa, which denotes the investments in additional annual injection capacity. Variable costs of CO2 storage are considered negligible as they sum up to less than 7-8 % of the overall storage costs (see section5.3.3 for further explanations).
All decision variables have to be non-negative (see Equation 2). Additionally, the number of pipelines to be constructed on one route inv_fijda are discretized (see Equation 3).
In the objective function each decision variable is multiplied by its respective cost factor. Eij is a distance matrix indicating whether two nodes i and j can be connected directly. If possible, the values of the matrix give the distances between i and j in kilometers. Scaling is easily done by varying the distance between nodes and their number and the spatial focus can range from regional to world-wide depending on research question and existing data sources.
The model is restricted by:
푥푃푎 + 푧푃푎 = 퐶푂2푃푎 ∀푃, 푎 (5.5)
Equation 5 defines that a facility’s CO2 stream can be treated in two ways, or a mixture of it: CO2 emissions can either be balanced with CO2 certificates (zPa), or the CO2 can be cycled through a capture system (xPa). Note that even if the entire CO2 stream is treated in the capturing facility (i.e. xPa= CO2) a fraction of (1-capt_rate) · xPa is released into the atmosphere and needs to be balanced with CO2 certificates (c.f. equation 1).
∑ 푓푖푗푎 − ∑ 푓푗푖푎 + ∑ (푚푎푡푐ℎ푃푃푗 ⋅ 푥푃푎 ⋅ 푐푎푝푡푟푎푡푒) − ∑ (푚푎푡푐ℎ푆푆푗 ⋅ 푦푆푎) = 0 ∀푗, 푎 (5.6) 푖 푖 푃 푆
Equation 6 specifies the physical balance condition, which states that all flows feeding into a node j must be discharged from the same node. match_PPj declares whether producer P is located at node j, while match_SSj declares whether a sink S is located at node j. The amout of CO2 that is transported and stored through the system is equal to the amount actually captured at the respective facility (xPa · capt_rate).
푥푃푎 ≤ ∑(푖푛푣_푥푃푏) ∀푃, 푎 (5.7) 푏<푎
The capturing capacity of each producer P in period a is given in equation 7. Note that all terms in this inequality are decision variables, meaning that injection in period a can only happen if the capacity was expanded prior to period a.
104 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
푓푗푖푎 ≤ ∑ ∑(푐푎푝_푑푑 ⋅ 푖푛푣_푓푖푗푑푏) + ∑ ∑(푐푎푝_푑푑 ⋅ 푖푛푣_푓푗푖푑푏) ∀푖, 푗, 푎 (5.8) 푏<푎 푑 푏<푎 푑
The capacity restriction of pipelines in Equation 8 works similarly to Equation 7.
푦푆푎 ≤ ∑ 푖 푛푣_푦푆푏 ∀푆, 푎 (5.9) 푏<푎
Inequality 9 states that the annual injection rate of a storage facility S is limited to the sum of investments in annual injection capacity inv_ySb from previous periods b.
( ) ∑ 5 ⋅ 푦푆푎 ≤ 푐푎푝푠푡표푟푆 ∀푆 (5.10) 푎
Inequality 10 restricts the amount of CO2 injected into reservoir S to its overall physical capacity. The multiplication by 5 resembles the amount of years per period a. Planning, licensing, and optimal routing of pipelines is ensured via Equation 11 where max_pipe is the maximum number of pipelines that can be built on a licensed route. The model is solved in the General Algebraic Modeling System (GAMS) using the CPLEX solver.
∑(푖푛푣_푓푖푗푑푎) ≤ 푚푎푥푝푖푝푒 ⋅ ∑(푝푙푎푛푖푗푏) ∀푖, 푗, 푎 (5.11) 푑 푏<푎
5.3 Application of the model for Europe and used data
5.3.1 CO2 emission sources Our European emission data covers the EU27 plus Switzerland and Norway. It includes industry facilities from iron and steel production, the cement and clinker production as well as oil refineries. Furthermore waste-, natural gas-, lignite- and coal-fueled power plants that emit more than 100,000 tCO2 per year are included. Facilities below this emission level are considered too small to justify the investment into capture, transport, and storage. Data on the average annual CO2 emissions of individual plants, location and age are taken from Platts (2011) and EEA (2011). We assume a lifetime of 55 years for lignite and hard-coal plants and 40 years for natural gas (NGCC) plants. Industrial facilities are assumed to be reconstructed with the same characteristics and on the same site once their economical lifetime ends. Projections on new power plant capacity installation are taken from VGB Power Tech (2011), covering 66 GW of NGCC, 7.6 GW of lignite and 35 GW of coal plants. Due to capacity aging and scrapping of old plants, this results in a decrease in fossil fuel capacity until 2050.
The total number of implemented emission sources in 2010 totals 2725 facilities, with emissions of
2.122 GtCO2 annually. These divide into 1476 (1.527 GtCO2/a) fossil fueled power plants and 1249
105 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
(0.595 GtCO2/a) industrial facilities. The graphical distribution of the included point sources is shown in Figure 5.2.
The CCTS investment costs for the sectors considered in this chapter are presented in Table 5.1.83 Costs estimates for the power generation sector are available from various sources (IPCC 2005; Tzimas 2009; ZEP 2011a; Finkenrath 2011; WorleyParsons and Schlumberger 2011). They all share the same general trend of lower capital cost for coal-fired generation compared to gas-fired power plants when calculated in € per tCO2. In the more recent studies a great share of the variation in the cost figures is attributed to changing raw material prices and different assumptions on the risk premium attributed to this immature technology. Costs for industry capture are gaining increased attention (Öko-Institut 2012; see e.g. Kuramochi et al. 2012). Rubin et al. (2007) examines learning rates of different climate protection technologies and estimates learning rates for carbon capture that we apply to our data from 2020 onwards.
Table 5.1: Investment costs for capture facilities (in €/tCO2pa). . (dimensioning of capturing sytem) Technology 2010 2020 2030 2040 2050 Coal 150 150 139 119 93 Lignite 116 116 107 92 72 NGCC 275 275 255 218 171 Cement 135 135 125 107 84 Iron and Steel 117 117 108 93 73 Refineries 210 210 195 167 131 Source: Own calculation based on various sources (Tzimas 2009; Ho, Allinson, and Wiley 2011; Öko-Institut 2012; Rubin et al. 2007).84
Table 5.2: Variable costs for CO2 capture (in €/tCO2). Technology 2010 2020 2030 2040 2050 Coal 32 32 31 31 31 Lignite 29 29 29 29 28 NGCC 47 47 45 44 44 Cement 17 17 17 17 17 Iron and Steel 16 16 16 16 16 Refineries 47 47 45 44 44 Source: Own calculation based on various sources (Tzimas 2009; Ho, Allinson, and Wiley 2011; Öko-Institut 2012; Rubin et al. 2007).
83 The depicted costs for CO2-capture do only cover the costs for the capturing unit itself, i.e. similar to retrofitting costs to an existing facility without CO2-capture. Overall system costs may vary depending on different generation types (power plants) or industrial facility. 84 Typically, investment and operating costs for CO2 capture are given in terms of MW and MWh, respectively. These figures refer to a specific capture rate (i.e. when making this investment one is able to capture a portion of the CO2 otherwise emitted into the atmosphere). The basic unit of the CCTS-Mod is tCO2. Thus we converted the figures accordingly to arrive at a per tCO2 based figure.
106 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Figure 5.2: CO2 emission sources and storage potential The variable costs of capture have two components: Variable costs of operation and maintenance and an energy penalty for additional energy input needed for the capturing process. Total variable costs are given in Table 5.2. For coal and lignite plants we apply the post-combustion capture technology. We assume the oxy-fuel process for the iron and steel and the cement sectors as proposed by Öko- Institut (2012). In this case, the variable costs of capture are mainly driven by the price for the electricity needed for the air separation unit. We assume a fixed price of 70 €/MWh, which remains constant. In refineries, we assume post-combustion capture. Due to the low CO2 concentration in the flue gas and the high diversity of the fuels and processes used in refineries, variable costs are comparable to those in natural gas plants (Ho, Allinson, and Wiley 2011).
5.3.2 CO2 transport Pipeline transportation is commonly considered as the most economically viable onshore transport solution that can carry the quantities emitted by large-scale CO2 sources. Onshore transport faces few technological barriers due to experience in the gas and oil sector and the CO2 industry for CO2-EOR in the USA. CO2 pipelines represent a typical network industry and are characterized by high upfront, sunk investment costs. Variable costs are comparatively insignificant and primarily include expenditures for fuelling compressors.
107 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
According to Heddle et al. (2003), right of way (ROW) costs account for four to nine percent of total gas pipeline construction costs depending on the diameter of the pipe, which we used to derive our values of the plan parameter. ZEP (2011b) presents a comprehensive study on CO2 transportation costs for different setting of transport networks. Calculated transport costs in € per tCO2 range from 2 to 20 depending on the network setting. Associated capital costs range between €0.08 and €0.15 per tCO2 and kilometer of pipeline. Topographic features, such as mountains or densely populated areas, are often neglected in studies as they need additional data and increase the computing time. Including such features, however, would lead to a strong increase of the transport costs or even infeasibilities of some projects (e.g. due to mountain ranges).
To account for the uncertainty associated with topographic features, public resistance, and environmental concerns as uncertain utilization rates we employ a higher value of 0.087 € per tCO2, cm of pipeline diameter, and km of pipeline. Economies of scale associated with CO2 pipeline transport pipelines are depicted through the five possible diameters with associated annual transport capacity (see Table 5.3).
Table 5.3: Investment cost by pipeline diameter and respective annual transport capacity.
Diameter [m] Annual transport capacity [MtCO2/ a] Investment costs [per tCO2 and km] 0.2 6 0.29 0.4 18 0.19 0.8 71 0.10 1.2 174 0.06 1.6 338 0.04 Source: Own calculations based on Ainger et al. (2010) and IEA (2005).
For operation and management (O&M) costs, ZEP (2011b) give values of €0.005 to €0.01 per tCO2 per kilometer. IEA (2005) arrive at similar operation costs varying between €0.01 and €0.025 per km per year depending on pipeline diameter and total pipeline length, including costs for booster stations; we thus use a value of €0.01 per year per km per tCO2 transported. Including the flow-dependent cost component ensures the shortest possible routing for the CO2. Planning and development (P&D) costs include ROW costs, land purchase and routing costs which occur only for the first pipeline built on a certain route. This leads to the construction of pipelines along corridors.
5.3.3 CO2 storage
Data on CO2 storage is difficult to come by and verify. Using available data, we derive our own estimates of location and capacity of the European on- and offshore storage. The exact location of the storage fields is being modeled as closely to the geological formation as possible. Various sources are used to get data for the UK and for offshore storage beneath the North Sea (M. Bentham 2006; M. S. Bentham, Kirk, and Wiliams 2008; Brook et al. 2009; Hazeldine 2009). Greenpeace (2011) give good estimates for storage potential in Germany, while Radoslaw et al. (2009) focus on Poland. The feasibility study for Europe-Wide CO2 Infrastructure from the European Commision (Ainger, Argent, and Haszeldine 2010) and the Geo Capacity (2009) project are used to estimate storage potential when no more accurate country specific study was available to the public. These studies, however, only grant public access to storage data on a 50x50km grid. This means that some of these formations
108 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
might consist of several smaller neighbouring aquifers. The example of Germany shows that the majority of the aggregated storage potential can actually be found in small reservoir of 50 Mt or less (Greenpeace 2011). The exploration of such small reservoirs is uneconomical, given a bad ratio of investment costs and exploitable storage capacity. The overall storage potential of Europe is thus over-estimated in these scenarios due to the lack of more detailed information. The total storage capacity is set to 94 Gt, spread among 41 Gt saline aquifers and 3 Gt depleted gas fields onshore and, offshore, 30 Gt saline aquifers and 20 Gt depleted gas fields (see Table B.2 in the Appendix).
According to Heddle et al. (2003) costs for CO2 storage are determined by factors including: type of storage facility, storage depth, permeability, number of injection points, injection pressure, etc. Therefore, total storage costs vary signifcantly in different studies (RECCS 2010). A characteristic value for a storage project is the sum of costs per injection well including site development, drilling, surface facilities, and monitoring investments for a given annual CO2 injection rate. Storage investments exhibit a strong sunk cost character and according to IEA (2005) variable costs total only seven to eight percent. Therefore, we implement storage costs on a total costs basis (see Table 5.4). A more recent estimate of storage costs from IEA GHG and ZEP (2011) examining different settings and uncertainties on technological and regulatory issues arrive at figures similar to those presented above.
Table 5.4: Site development, drilling, surface facilities and monitoring investment cost for a given annual CO2 injection rate per well. Natural gas field Saline aquifer Technology Onshore Offshore Onshore Offshore Drilling length [m] 3000 4000 3000 4000 Well injection rate [(MtCO2/a)] 1.25 1.25 1 1 85 Corrected well injection rate [MtCO2/a] 0.4 0.4 0.33 0.33 Drilling costs [€ per m] 1750 2500 1750 2500 Investment in surface facilities [M€] 0.4 25 0.4 25 Monitoring investments [M€] 0.2 0.2 0.2 0.2 Wells per location 6 6 6 6 Total drilling costs [M€] 5.25 10 5.25 10 Total capital costs per well [M€] 5.6 14.5 5.6 14.5 O&M and monitoring costs [%] 7 8 7 8 Source: Own calculations based on IEA (2005).
One option that is said to improve the economics of CO2 storage and CCTS in general is CO2-EOR (IEA and UNIDO 2011). The technology is increasingly used in the USA and Canada (MIT 2016) and might also be an option to provide additional investment incentive for CCTS projects in Europe. Studies look intensively into the interaction of these two technologies. Some regional studies on the UK and Norwegian potential (Kemp and Kasim 2013; Klokk et al. 2010) as well as larger scope studies (e.g. on the North Sea region (Mendelevitch 2014) or the US (Davidson, Dahowski, and Dooley 2011)) are also available. In general, it is up to future research to determine whether the combination of the two technologies can still be considered as CO2 abatement, when taking into account the emissions
85 According to Gerling (2010), an annual injection rate of 300,000 to 400,000 tCO2 per well is more realistic for most formations.
109 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
from additionally recovered oil and assessing different injection strategies (see e.g. ARI and Mezler-
Consulting 2010). For our approach we do not consider CO2-EOR as a storage option (see Chapter 6).
5.4 Different scenarios and their results analyzing political and geological uncertainties The level of uncertainty about the size and configuration of the pipeline network emanates from the uncertainty about future carbon policies, the level of deployment of renewable energy technologies, as well as the suitability and usability of geological formations to store captured CO2. Different scenarios are implemented with a linear increase in CO2 prices from 15 €/t in 2010 until 2050. For the base case the CO2 certificate price increases from 15 €/t in 2010 to 75 €/t in 2050. Additionally, we define a scenario with a higher (100 €/tCO2) and a lower (50 €/tCO2) CO2 price in 2050. We do not implement a correlation between CCTS deployment and the price for CO2. We also consider the possibility that onshore storage may not be possible in Europe, due to technical, political, or whatever other reasons. In that case, storage would need to take place offshore, mainly in the North Sea, and the total storage potential would be significantly reduced, from 94 Gt (on- and offshore) to only 50 Gt. The respective scenario key assumptions are shown in Table 5.5.
Table 5.5: Key scenario assumptions.
Scenario CO2 price in 2050 [€/tCO2] Storage {on;offshore} CO2 storage capacity [Gt] Ref75 75 on and offshore 94 Off75 75 offshore 50 On50 50 on and offshore 94 Off50 50 offshore 50 On100 100 on and offshore 94 Off100 100 offshore 50 Source: Own depiction.
5.4.1 Reference scenario: certificate price increasing to 75 €/tCO2 in 2050
5.4.1.1 On- and offshore storage Our Reference scenario simulates the cost-optimal deployment of a European CCTS infrastructure for the period 2010-50 given a CO2 certificate price starting at €15 in 2010 and rising to €75 in 2050. Point source emissions, storage sites and potential pipelines are mapped on a spherical grid covering Europe. The distance between two neighboring grid nodes is two degrees (on average about 200 km).
In this Reference scenario, 758 Mt of CO2 emissions are captured, transported, and stored annually through CCTS in 2050. CCTS implementation begins in 2020 with the first investments. The capturing process starts five years later in both the iron and steel as well as in the cement sectors. CCTS infrastructure gradually ramps up from 2020 to 2040 (see Figure 5.3). At first, the industrial facilities with lower capturing costs situated close to potential storage sites are the predominant users of CCTS.
Industrial CCTS penetration reach a capturing rate of 370 MtCO2 per year in 2050. With rising CO2 prices CCTS becomes a more attractive abatement option for the power sector. The annual rate of stored CO2 from power generation reaches 390 Mt in 2050 (see Table 5.6 for an overview of key results).
110 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Figure 5.3: Storage by sectors in MtCO2, Ref75 Over the 40 year modeling time horizon, total investment costs along the CCTS value chain sum up to
€240 bn. Given a total quantity of avoided emissions of 15.8 GtCO2, this breaks down to investment cost of €15.3 per tCO2 avoided. Total variable costs sum up to €515 bn, or €33 per tCO2 avoided. Although this number may appear low, we note that most capture occurs in the industrial sector in the early years. The costs of the capturing process hereby comprise around 90% of the total costs while the transport and storage only have minor impacts assuming an optimal grid and storage planning (see Figure 5.4).
We note that under the applied CO2 price path, CCTS is an option primarily for countries with a regional proximity between CO2 intensive regions and storage sites. The technology is mostly implemented by Poland, Germany, the Netherlands, Belgium, France, and the UK. Moreover, we find no interconnected, transnational transportation network (see Figure 5.5). As industry facilities will be the first-movers, they drive the layout of the pipeline network.
111 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Figure 5.4: Infrastructure investment and variable costs in €bn, Ref75
Figure 5.5: CCTS infrastructure in 2050, Ref75
112 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
5.4.1.2 Offshore storage only Due to longer transport distances and more expensive storage, this sub-scenario leads to a deployment of CCTS on a lower level compared to the Reference scenario in section 5.4.1. Over the 40 year modeling time horizon, total investment cost along the CCTS value chain total €145 bn. Capture investment occurs in two waves, the first in industry in 2025 and the second in the power sector in 2040 (see Figure 5.6 and Figure 5.7). This is a delay of 5 years compared to the reference scenario (see Figure 5.3 and Figure 5.4). Given a total quantity of avoided emissions of 7.5 GtCO2, this breaks down to investment costs of €19.4 per tCO2 avoided, an increase to the reference scenario of 22%. Total variable costs sum up to €266 bn, or €35.4 per tCO2 avoided. With only a slightly higher participation of the power sector, this increase in the average variable costs of CO2 abatement compared to the reference scenario is explained by longer transport distances and more expensive offshore storage.
Figure 5.6: CCTS infrastructure in 2050, Off75
113 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Figure 5.7: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, Off75 We also note a lower participation of the cement sector in CCTS whereas capture in the iron and steel sector remains at the same level. This is explained by higher capture costs in the cement industry (see Table 5.1), but also by the geographical distribution of industrial facilities: while emitters in the iron and steel sector are generally located close to the coast, cement kilns are often located close to inland mining facilities. Thus, a possible strategy could be to form regional clusters that could more easily benefit from economies of scale in transport.
Results of an offshore only scenario for Germany on a much higher resolution (distance between nodes only 50 km) show a greater drop in CCTS deployment compared to the results presented in this chapter (Oei, Herold, and Tissen 2011). The primary reason for that is the distance of 200 km in between nodes which strongly overestimates economics of scale in transportation since many emitters are grouped and also often set closer to storage sites than in reality. Yet the distance, and therefore the total number of nodes for a modeling region, is limited by computational runtime which increases exponentially with the number of nodes. The scenarios in this chapter, which use 460 nodes, require a runtime between 48 and 72 hours on a machine with 8 cores and 30 GB RAM86.
5.4.2 Certificate price increasing to 50 €/tCO2 in 2050
5.4.2.1 On- and offshore storage
Earlier results of the CCTS-Mod focusing only on Germany show that an increase in the CO2 certificate price to €50 per tCO2 leads to an application in industry only (Oei, Herold, and Tissen 2011). Those findings are confirmed by the CCTS-Mod on the European level as well. The lower costs of capture again lead to investments in the steel industry first, followed by the cement industry five years later (see Figure B.9 in the Appendix).
The CCTS technology primarily remains an abatement option for large industry clusters with a regional proximity to storage sites in Northern Europe. This excludes small and mid-scale facilities in the
European hinterland. However, with a total storage of 5.6 GtCO2 over the next 40 years, this scenario shows the potential for CCTS in the iron and steel and cement sector even at a low CO2 certificate
86 2x Intel Xeon X5355 2.66 GHz Quad-Core, 8 MB Cache
114 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
price. Investment cost along the CCTS value chain totals €81.4 bn. This leads to average investment costs of €14.6 per tCO2 avoided. Total variable costs sum up to €134 bn, or €24 per tCO2 avoided.
Figure 5.8: CCTS infrastructure in 2050, On50
5.4.2.2 Offshore storage only
In the case of offshore storage only the total storage is reduced to 2.1 GtCO2 over the next 40 years at total investment cost of €40 bn. This leads to average investment costs of €18.5 per tCO2 avoided. Several industrial facilities are located far from any offshore site and thus do not invest in CCTS. They are relatively scattered and cannot form large enough emission clusters to benefit from economies of scale with transporting the CO2 over longer distances. The total variable costs sum up to €58 bn, or
€26.4 per tCO2 avoided. Average variable costs are much lower in case of the certificate price remaining below €50 as the high cost power sector is not investing in the CCTS technology.
This scenario highlights the importance of available onshore sinks, especially for the promotion of the
CCTS technology at moderate CO2 prices. However, the debate on onshore storage in several European countries (e.g., the Netherlands and Germany) indicates that this storage option could be ruled out by regulation.
115 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
5.4.3 Certificate price increasing to 100 €/tCO2 in 2050
5.4.3.1 On- and offshore storage
This scenario results in a total storage of 24.7 GtCO2 over the next 40 years, a significant increase compared to the reference scenario. The same is true for the investment costs along the CCTS value chain, which increases to €380 bn. This leads to average investment costs of €15.4 per tCO2 avoided.
Total variable costs increase to €929 bn, or €38 per tCO2 avoided. This can be explained primarily by the higher participation of the power sector in CCTS.
5.4.3.2 Offshore storage only
This scenario results in a total storage of 19 GtCO2 over the next 40 years. The investment costs along the CCTS value chain add up to €359 bn or an average of €18.7 per tCO2 avoided. The total variable costs are €796 bn or €41.5 per tCO2. The cost increase is based on longer transport distances and the greater participation of the power sector.
Figure 5.9: CCTS infrastructure in 2050, On100 Table 5.6 provides a summary of the scenario results, in terms of the required pipeline network, total stored emissions, the share of emissions that originate from industrial sources, remaining storage potential, as well as total costs for CCTS (fixed and variable).
116 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
Table 5.6: Overview of scenario results Pipeline Stored Storage left Network Emissions Origin from in 2050 CCTS inv. CCTS var. Scenario [km] [GtCO2] industry [%] [GtCO2] costs [bn] costs [bn] Ref75 20,400 15.8 63 78.2 240 515 Off75 9,800 7.5 65 42.5 145 266 On50 6,600 5.6 100 88.4 81.4 134 Off50 4,300 2.1 100 47.9 40 58 On100 23,600 24.7 53 69.3 380 929 Off100 37,400 19 57 31 359 796 Source: Own calculation.
5.5 Conclusion: the future of a CCTS roll-out in Europe The role of CCTS in future decarbonization portfolios is highly uncertain. Part of this uncertainty is due to a lack of objective information and independent economic analysis. To improve the situation, we develop a model suggesting optimal strategies for deploying a carbon capture, transport, and storage infrastructure. The model integrates technical details, focusing on a simple decision rule on whether
"to capture or not to capture": emitters can pay a given CO2 price, or else engage into CCTS to abate their CO2; the model will minimize the costs of both, purchase of CO2 certificates and CCTS- infrastructure.
With respect to the existing literature, we include new features into the model, such as the explicit recognition of planning costs, as well as the option to combine CCTS in the industry and the electricity sectors. The model suggests that under certain assumptions, CCTS may contribute to the decarbonization of Europe’s industry and energy sectors. However, only if the CO2 certificate price rises to €75 by 2050 and sufficient CO2 storage capacity is available both on- and offshore, will CCTS have the potential to play a role in future energy technologies.
Our results indicate that given an increase in the CO2 certificate price up to 50€/tCO2 in 2050, deployment will be limited only to industrial applications in the iron and steel as well as cement sectors. The infrastructure will remain regional without Europe-wide integration. However, European cooperation could still be of benefit in areas where emission sources and sinks are divided by national borders and for offshore storage solutions.
In all scenarios, industry plays an important role as a first mover to induce deployment. A decrease of available storage capacity or a lower increase in future CO2 certificate prices could significantly reduce the role of CCTS as a CO2 mitigation technology, and especially its role in the decarbonization of the electricity sector. We also observe an initial decline in per unit expenditures for CO2 transport in scenarios with broad CCTS utilization, due to economies of scale. In later periods this effects is, however, partly offset by increasing transport distances due to the development of more distant storage resources, once the close and cheap ones are exhausted.
In this context, the storage capacity left at the end of the modeling horizon in 2050 might also be misleading, at first sight. In a post-2050 horizon, cheap storage resources are used up and more distant and costly storage sites will need to be developed. On the other hand, experience gained with developing and operating CO2 storage sites can also modulate this costs escalation or even lead to
117 Chapter 5: Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe
overall cost reductions. A quantification of the different effects is up to future research. Another aspect often being neglected is the need for reserving affordable storage options when hoping for negative
CO2 emissions through Biomass–CCTS in some decades. Such and other competing concepts for utilizing underground resources (compressed air storage, natural gas / oil storage, geothermal power and / or heat recovery) make it difficult to estimate the remaining usable storage potential for CCTS.
Given continued social and political opposition to onshore storage, CO2 abatement by means of CCTS, seems only viable with respect to offshore storage. We suggest that policy-makers give first priority to CCTS for coastal areas and small industrial sites where CO2 transport does not require intensive infrastructure investments to prove the technology’s viability, especially in the industry sector. The additional costs of longer pipelines and higher costs for storage development in all offshore scenarios lead to a delay in the CCTS implementation of at least five years. However, in reality, this could well be offset by shorter planning processes if the public accepts offshore transport and storage.
Note that our model runs assume a single planner basing its investment decision on full insights into remaining storage capacities, the future CO2 price development and actions of all other emitters. The outcomes therefore overestimate the potential for CCTS investment. The key uncertainty of the model is the CO2 certificate price; its influence on the CCTS-deployment can be seen in the different scenario runs. The variable capturing costs are the second biggest uncertainty of the model and are mainly driven by the electricity price. An increase of these capturing costs would slow down the deployment of CCTS. Transport costs sum up to 10%, while storage costs lie below 5% of the overall CCTS-costs in all onshore scenarios. These figures, however, nearly double in the offshore scenarios. Mapping emission sources and sinks to nodes also affects the results, mainly by underestimating the necessary transport infrastructure and overestimating economies of scale. Future research should focus on advanced modeling techniques reducing model runtime to enable a European model run with a higher resolution.
Our scenario analysis underlines that the future development of an integrated CCTS infrastructure is highly sensitive to assumptions regarding the future CO2 certificate prices and the availability of storage resources. If CCTS is to become a cornerstone of a future low-carbon industry and power generation sector policy makers have to commit to clear and reliable targets regarding the future CO2 prices, or provide alternative long-term investment incentives. Getting the industry sector back into the CCTS debate will help to change the public opinion towards CCTS, when confronted to the lack of alternatives. Based on the persistent experience of canceled and postponed CCTS demonstration projects and reluctant institutional and private investors the authors doubt that CCTS will become the integrated pan-European industry once envisioned by EU-level policy makers.
118 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Chapter 6
EUROPEAN SCENARIOS OF CO2 INFRASTRUCTURE INVESTMENT UNTIL 2050
6.1 Introduction: CO2-Enhanced oil recovery keeps the mirage alive Carbon Capture, Transport, and Storage (CCTS) was originally seen as a central element for decarbonized electricity systems, worldwide (e.g. IEA 2010).87 The International Energy Agency (IEA) consequently underlined its importance with a 20% contribution to achieving emission reduction goals and 40% cost increase for decarbonization in its absence (IEA 2012a). Even higher cost increases of 29-297% are estimated by the IPCC (2014) for reaching the 2°C target. Estimates for the European energy system projected 77 (IEA 2012a) to 108 GW (EC 2011) of power generation capacity to be equipped with CCTS and a CO2 transport network of over 20,000 km by 2050 (JRC 2011). The reality, however, is in great contrast to these expectations, as documented in a special issue by Gale et al. (2015b) commemorating the 10th anniversary of the IPCC (2005) special report. Not a single full-scale CCTS project with long-term geological storage has yet been realized worldwide (GCCSI 2014). At the same time, CO2 transport infrastructure projects have been removed from the list of critical infrastructure projects of the EU (EC 2013a). Furthermore, the London Protocol still prohibits the movement of CO2 across marine borders for the purposes of geological storage (GCCSI 2014). Facing these adverse developments, academia as well as technical reports have become more balanced or even critical with respect to CCTS deployment (Hirschhausen, Herold, and Oei 2012b).
The gridlock in the deployment of CCTS can be partly explained by low EU Emissions Trading System
(EU-ETS) CO2 prices which have remained in the range of three to eight €/tCO2 since the start of the third trading period in 2013. These low prices – with little hope for significant rise in the coming years (Hu et al. 2015) – give insufficient incentive for investment into mitigation technologies such as CCTS. Investment costs for renewables, in contrast, have profited from high learning curves and have become a much cheaper abatement option. Even additional financial schemes such as the European Energy Program for Recovery (EEPR) proved unsuccessful in enabling projects (GCCSI 2014). The New Entrance Reserve (NER300) program, originally designed to provide up to €9 bn worth of funding renewables and CCTS projects, has a budget of only €1.5 bn due to its revenue being based on the
87 This chapter is forthcoming in The Energy Journal. A previous version has been published as Resource Markets Working Paper WP-RM-36 at University of Potsdam (Oei and Mendelevitch 2013). It is joint work with Pao-Yu Oei. Roman Mendelevitch and Pao-Yu Oei jointly developed the model and its implementation in GAMS. Pao-Yu Oei had the lead in analyzing the political setting for CCTS in the EU. Roman Mendelevitch had the lead in collecting data on CO2-EOR, and analyzing the results. The writing of the manuscript was executed jointly.
119 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
sale of 300 million CO2 allowances. As a result, none of the 12 CCTS projects that applied for funding in the first round were supported (Lupion and Herzog 2013). In July 2014 the second round of the NER300 granted €300 million to the UK White Rose CCS Project. Meanwhile, the original project timeline was pushed back by two years, aiming at completion only in 2020.88 The project outcome became even more unlikely when one of the main investors decided to withdraw in September 2015.89 Martinez Arranz (2015) identifies various blind spots in the EU demonstration programs, as Europe, in comparison to other regions, is a relatively resource-poor but advanced economy. He therefore recommends a stronger focus on the industrial use of CCTS as well as other non-CCTS mitigation possibilities in the power sector.
In Europe, the directive on the geological storage of CO2 (so-called "CCS Directive") is the central regulatory element intended to govern the process of CCTS commercialization.90 However, this directive limits the scope of underground storage to non-commercial sizes, which is insufficient for large scale projects.91 Although focusing on the storage part of the technology chain, the Directive also requires “CCTS readiness” for new fossil generation capacities. Lacking a clear definition for this “readiness”, the Directive leaves space for interpretation. A review process conducted by the Directive in 2014 highlights the need for running CCTS demonstration projects in Europe. In particular, it criticized the lack of progress of CCTS for industrial applications such as steel or cement facilities, which contribute up to a quarter of the world’s energy-related CO2 emissions. One option that many stakeholders requested during the review process was a successor NER300 scheme starting in 2020 onwards to support future projects.92
Complementary to price incentives, CCTS in some countries is promoted via climate-oriented regulation or in combination with enhanced oil recovery (CO2-EOR) projects. The introduction of Emissions Performance Standards (EPS) in the UK, Canada and the U.S. restrict the annual amount of CO2 emissions per installed unit of generation capacity and thereby the operation of new coal power 93 plants without CO2 capture . Using captured CO2 for EOR purposes contradicts the idea of long-term geological storage but significantly improves the cost effectiveness of a CCTS project. Successful projects like Boundary Dam in Saskatchewan, Canada (in operation since October 2014), as well as the majority of upcoming projects in 2016-17 (e.g. Kemper County Energy Facility and Petra Nova
Carbon Capture Project in the US) are associated with CO2-EOR. Little progress, however, is seen in
88 See EC (2014a). European Commission - Press release - Climate action: Commission uses polluters’ revenues to fund clean energy projects across Europe, as well as Szabo, M. (2014): EU pushes back deadlines 2 years in green energy funding scheme. Reuters UK. http://uk.reuters.com/article/2014/10/15/eu-renewables-ner- idUKL6N0SA48320141015 (03/08/2016). 89 Drax pulls out of £1bn carbon capture project http://www.bbc.com/news/business-34356117 (16/10/2015). 90 See EC (2013b). Draft proposal for a Directive amending Nuclear Safety Directive IP/13/532 (Brussels, Belgium: European Commission). 91 Triple EEE Consulting (2014). CCS Directive Evaluation. 92 Triple EEE Consulting, Ricardo-Aea, and TNO (2015). Support to the review of Directive 2009/31/EC on the geological storage of carbon dioxide (CCS Directive) (The Netherlands). 93 The UK has introduced an Electricity Market Reform (The Parliament of Great Britain 2013), where one of the four pillars builds on EPS benchmarked against gas-fired electricity generation,; similarly, the US (EPA 2012, final rule pending for submission to the federal register since 05.08.2015) and Canada (Parliament of Canada 2012) have introduced EPS for new electricity generation units.
120 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
the EU as only a few riparian states of the North Sea are capable of CO2-EOR projects. Nevertheless, the EU framework for climate and energy still aims at a commercial CCTS deployment by the middle of the next decade (EC 2014a).
This chapter presents a model analysis and interpretation of the potential role of CCTS to support the EU energy system transition to meet the emission reductions goals that are consistent with an international goal of staying below 2°C of global warming. Our hypothesis is that CCTS – contrary to the dominant belief until recently – will at best be a niche technology applied in regions with highly conducive conditions, e.g. parts of the North Sea, but that due to its cost disadvantages and recent setbacks in many EU countries, will not contribute significantly to overall EU decarbonization.
Moreover, the discrepancy between locations of CO2 emissions and the availability of potential CO2 storage sites call for regional cooperation, which is vital to the economies of scale associated with CO2 transport infrastructure. Few papers currently address this issue (e.g. Geske, Berghout, and van den Broek 2015b) or try to find solutions to the resulting coordination game (e.g. Massol, Tchung-Ming, and Banal-Estañol 2015). We quantify scale economies of different CCTS network coverages.
Section 6.2 provides a non-technical description of our CCTS-Mod; a multi-period, scalable, mixed integer framework calculating beneficial investments in the CO2-chain (capture, transport, storage). Section 6.3 presents the outcome of the European-wide results. We find no role for CCTS in the 40% mitigation scenarios. In the 80% mitigation scenarios, some CO2-intensive industries might start to abate, followed by the energy sector at a high CO2 price (above 100 €/tCO2). We consider this scenario unlikely, because most of the countries involved have already given up CCTS as a mitigation option, e.g. Germany, Poland, France, and Belgium. Section 6.4 focuses on an alternative driver for
CO2-abatement through CO2-EOR. We find that for North Sea riparian countries that have not given up on CO2 capture, mainly the UK and Norway, the use of CO2-EOR might be an economical option, depending on oil prices and prices of CO2 certificates. Once CO2-EOR resources are fully exploited, further CO2 capture activity is solely incentivized by CO2 certificate prices, which must cover at least the variable costs but also potential new investment costs. Also, the speed and extend of the deployment is highly dependent on assumptions for initial technology costs and learning effects. Section 6.5 concludes by analyzing the chances for a regional vs. European-wide CCTS application depending on the availability of CO2-EOR and other storage potentials.
6.2 Model, data, and assumptions
6.2.1 The model CCTS-Mod For our numerical analysis, we use the “CCTS-Mod” (see Chapter 5 for a detailed model description). The model is a multi-period, scalable, mixed integer model coded in GAMS (General Algebraic Modeling Software) and solved with a CPLEX solver. For each power plant or industrial facility covered in our input database (see section 6.2.2), an omniscient planner decides on whether to invest into a CCTS chain or to buy CO2 certificates. The model decides in favor of CCTS whenever the net present value of CO2 certificates required to cover emissions during the model horizon (2055) is higher than the net present value of all costs related to CCTS.
121 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
In this case, investments into a capture unit facing respective capital costs have to be made. It takes five years after the investment decision before the capture unit becomes operational. Whenever a facility is used to capture CO2, variable costs are induced. The capture rate is capped at 90%. CO2 capture has to be balanced with CO2 transport and storage. Again, respective infrastructure investments have to be made taking into account a construction period of five years. Capital costs for transport cover right of way (ROW) costs and other investment cost parameters. If a new pipeline is constructed along a route that is already developed, ROW costs do not apply. This ensures that transportation routes are bundled in corridors, which is consistent with practices for laying natural gas or crude oil pipelines. The construction of a pipeline is a binary decision with discrete pipeline diameters and associated throughput volumes. CO2 storage is again subject to a five year construction period and has associated variable and capital costs.
Figure 6.1: Decision tree of the model CCTS-Mod with the option of CO2-EOR. A refined version of the model which is used for the model runs of this chapter includes the option to use captured CO2 for enhanced oil recovery. CO2-EOR storage is associated with additional investment and variable costs for equipment and operation, but generates revenue from oil recovered with each ton of CO2 stored. A simplified decision path of the CCTS-Mod is illustrated in Figure 6.1.
122 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
The model is based on literature on CCTS infrastructure models, developed by Middleton and Bielicki (2009) and Morbee et al. (2012).
The motivation behind the model is to provide an estimate on overall costs and needed infrastructure of a possible CCTS roll-out in Europe. Such figures are important from a central planner’s perspective and serve as bottom line for future cost estimates. The main drivers of the model are location and volumes of CO2 emissions, storage capacities, investment and variable costs at each stage of the
CCTS technology chain, and assumptions about future CO2 certificate and oil prices. The development of the CO2 price is thus exogenously given, independent from the CCTS deployment allowing for better comparison between the scenarios.94 Several uncertainties persist regarding the model: First, the cost-minimizing approach from a central planner’s perspective likely underestimates the actual costs of CCTS technology, as we assume perfect foresight, no market power by individual companies or sectors as well as a vertically integrated CCTS chain. Second, the model assumes the existence of certain technologies that have not been proven to work in practice on larger scales. The
“cost” estimates for CO2 capture and storage are especially uncertain, and most likely highly underestimated. The model also does not take into account the transaction costs of bringing the immature technology to implementation, to build infrastructure or to develop the storage sites. We also do not include costs associated with rising public opposition. However, total cost figures might be also overestimated, due to the limited utilization horizon of the infrastructure.
6.2.2 European data set Data was collected for the period between 2015 and 205595 and is based on a more detailed description of the cost data in Mendelevitch (2014). The scope of this study includes all members of the EU as well as Switzerland and Norway, and their respective Exclusive Economic Zones (EEZs). Data on location and emission volumes of refineries, steel and cement production facilities as well as coal- and gas-fired power plants is taken from the database developed in Chapter 5.
94 The exogenous CO2 certificate allows for no impact of CCTS investments on the certificate price. This limitation is necessary as such an analysis is currently only possible if technical restrictions - like non-linearities – are not considered (see e.g. Kanudia, et al. (2013)). 95 Note that model results for 2055 will not be interpreted. This last period is introduced to include an additional payback period and to allow for investment in 2050.
123 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Figure 6.2: Distribution of potential CO2 storage sites (left) and CO2 source (right) by type and volume in the data set. Source: Own illustration based Chapter 5 and Mendelevitch (2014).
124 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
The database assumes an economic lifetime of 40 years for gas-fired and 50 years for coal-fired power plants. Power generation facilities are supposed to be shut down without replacement after the economic lifetime is reached while industrial plants are assumed to be replaced by facilities with similar characteristics. It is unclear how emissions from industrial facilities and power plants will evolve in the future. In the electricity sector, a reduction of emissions is very likely due to rising shares of renewable energy sources. The model therefore includes projections for the closure of existing plants as well as the construction of new conventional power sources resulting in a reduction of overall CO2 emissions from the electricity sector. In the industry sector a reduction of emissions due to efficiency improvements, increased recycling shares, alternative industrial processes as well as relocation of industries away from Europe are to be taken into account. A lack of source specific data for the industry including e.g. information regarding construction date or planned new sites and relocations, however, make such estimates very complicated. The model therefore assumes a continuous flat profile in the industry sector that overestimates the potential for CCTS in the industry. The model hereby assumes that old industrial facilities are replaced with new ones at the same location, not taking into account strategic relocation due to CCTS costs at other locations. This underestimates the application of CCTS in the industry sector. We assume this latter effect to be relatively small as the
CO2 capture costs are the biggest share of costs and independent of the source location. The same database was used for location and capacities of potential storage in depleted hydrocarbon fields and saline aquifers. Data on location and volumes of CO2-EOR storage sites is taken from Mendelevitch (2014). Figure 6.2 illustrates the distribution of emission sources and their respective emission volumes for 2010 as well as the distribution of storage sites by type and their respective capacities. It depicts the fact that emission sources and storage sites are not equally spread across Europe. While the largest emission sources are located in the Rhine Area, the largest storage capacities can be found offshore in the UK and Norwegian EEZs.96 Denmark, UK and Norway are the only countries that have potential for CO2-EOR in their parts of the North Sea. Strong opposition in several European countries have formed against onshore CO2 storage. All scenarios in this chapter therefore only include the option of storing the CO2 in offshore fields.
6.2.3 Mathematical formulation The mathematical formulation is the same as presented in Chapter 5, except that Equation (5.1) is extended by one more item, namely potential revenue generated from CO2-EOR. The one period revenue is calculated when multiplying the underlying 표푖푙푝푟푖푐푒푆푎 by the amount of CO2stored via CO2- EOR, as shown in Equation (6.1).
96 The estimates for possible storage locations are based on studies which mostly offer data on a 50 x 50 km grid. Some of these formations, however, consist of several smaller neighboring aquifers. The exploration of small reservoirs is less economical, given a bad ratio of investment costs and exploitable storage capacity. The overall storage potential of Europe is thus overestimated in this chapter due to the lack of more detailed information.
125 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
1 (푦푒푎푟푎−푦푒푎푟푠푡푎푟푡) min ∑ [( ) 푥푃푎,푖푛푣_푥푃푎, 1 + 푟 푧 ,푓 ,푖푛푣 , 푎 푃푎 푖푗푎 푓푖푗푑푎 푝푙푎푛푖푗푎,푦푆푎,푖푛푣_푦푆푎
⋅ (∑[(푐_푐푐푠푃푎 + (1 − 푐푎푝푡_푟푎푡푒) ⋅ 푐푒푟푡푎) ⋅ 푥푃푎 + 푐_푖푛푣_푥푃 ⋅ 푖푛푣_푥푃푎 + 푐푒푟푡푎 ⋅ 푧푃푎] 푃
+ ∑ ∑ [퐸푖푗 ⋅ (푐_푓 ⋅ 푓푖푗푎 + ∑(푐_푖푛푣_푓푑 ⋅ 푖푛푣_푓푖푗푎푑) + 푐_푝푙푎푛 ⋅ 푝푙푎푛푖푗푎)] 푖 푗 푑
+ ∑[푐_푖푛푣_푦푆푎 ⋅ 푖푛푣_푦푆푎] − ∑ [표푖푙푝푟푖푐푒푆푎 ⋅ 푦푆푎])] 푆 푆=퐸푂푅
(6.1)
6.2.4 Assumptions for all scenarios Two key parameters drive the results of our model runs: On the one hand CCTS deployment is triggered by the CO2 certificate price path which governs the profitability of the CCTS technology in comparison to balancing CO2 emissions with purchased CO2 certificates. If in the long run, anticipated prices are higher than the costs of using the technology chain, then CCTS is employed. We use two possible price pathways generated by the PRIMES model (EC 2013b) which represent the outcomes of two sets of scenarios for climate change mitigation policy up to 2050 (see Table 6.1). The 40% scenarios include the EU 2020 targets as well as a 40% greenhouse gas (GHG) reduction by 2050 compared to 1990. The 80% scenarios are more ambitious including an 80% GHG reduction by 2050. All scenarios do not allow for emission trading across macro regions i.e. no trade between EU-ETS and RGGI or other schemes (but trade within macro regions, e.g. within the EU through a cap and trade system). They include moderate assumptions on efficiency gains and availability of nuclear and renewable energies (see Holz and von Hirschhausen (2013) and Knopf et al. (2013) for a detailed description of the underlying assumptions).
Table 6.1: CO2 certificate price path in the different scenarios. Scenario 2015 2020 2025 2030 2035 2040 2045 2050 Certificate price in 40% 14 17 27 37 45 52 52 52 €/tCO2 80% 18 25 39 53 75 97 183 270 Source: Knopf et al. (2013). The availability of storage capacity is the second decisive parameter. France, Germany and Belgium, in particular have their storage resources located mostly in onshore saline aquifers and depleted hydrocarbon fields. However, onshore storage is associated with significantly higher complexity of regulation and a higher number of stakeholders involved. The Global CCS Institute has performed a comprehensive assessment of CO2 storage readiness on a country level and come to the conclusion that Norway is the only European country currently ready for a wide-scale CO2 storage deployment (GCCSI 2015a). Germany, the Netherlands and the UK are the only countries that are at least ranked advanced. The assessment reveals a high correlation between a country’s ranking and the existence of an advanced hydrocarbon industry, and its dependence on fossil resources. Following long
126 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
debates, onshore storage was excluded as a storage option in Germany (Schumann, Duetschke, and Pietzner 2014, 2) 97, Denmark98, the UK and Netherlands (GCCSI 2012). Analogous developments are conceivable for other countries, leaving offshore storage as the only remaining storage option in Europe. Accordingly, none of the Europe-based large-scale integrated CCTS projects listed in the
Global CCS Institute database include onshore CO2 storage (GCCSI 2014). Therefore, in all presented scenarios, onshore storage capacity is not available, which reduces total available storage capacity from 94 GtCO2 to 50 GtCO2 in the European-wide scenarios and from 56 GtCO2 to 42 GtCO2 for the scenarios which focus on the North Sea region. As a consequence, France and Belgium lose most of their domestic storage potential. Despite a number of minor storage resources (1.2 GtCO2) in saline aquifers in the German North Sea, the situation in Germany is similar.
The resulting scenarios shown in Table 6.2 differ in their respective CO2 price path, the availability of storage potential (offshore with vs. without CO2-EOR) and geographical coverage (European-wide vs. the North Sea region vs. selected countries). Section 6.3 describes the European scenarios (EU_40% and EU_80%) while section 6.4 further analyzes regional scenarios (NorthSea_40%, NorthSea_80% and DNNU_80%).
Table 6.2: List of scenario assumptions.
Scenario Coverage CO2 price in 2050 Storage availability EU_40% Europe 52 €/t Offshore only EU_80% Europe 270 €/t Offshore only NorthSea_40% North Sea region 52 €/t Offshore only + CO2-EOR NorthSea_80% North Sea region 270 €/t Offshore only + CO2-EOR DNNU_80% DK, NL, NO, UK 270 €/t Offshore only + CO2-EOR
6.3 Results of the European-wide scenario analysis
6.3.1 EU_40% scenario
CCTS starts being deployed from the year 2035 onwards when the CO2 certificate prices pass the 40
€/tCO2 threshold. Nevertheless only a very small annual amount of around one MtCO2 is captured and stored in offshore hydrocarbon fields as well as saline aquifers. Hydrocarbon fields are the cheapest storage option when excluding CO2-EOR, but are not available in all locations. Four iron and steel factories in Norway and Estonia are the only emitters that invest in capture technology, benefiting from the lower variable and fixed costs assumed for this industry. The investing factories are located at the coast which leads to lower transport costs than for other industrial facilities. The overall costs sum up to €0.2 bn of investment costs and an additional €0.4 bn of variable costs until 2050.
6.3.2 EU_80% scenario
The increase of the CO2 price in the EU_80% scenario is higher than in the EU_40% scenario. The price increases gradually until a stronger rise kicks off in 2030, resulting in its final value of 270 €/tCO2
97 See Hirschhausen, C. von, Herold, J., Oei, P.-Y., and Haftendorn, C. (2012b). CCTS-Technologie ein Fehlschlag: Umdenken in der Energiewende notwendig. DIW-Wochenbericht 79, 3–9. 98 See Brøndum Nielsen, L. CCS | Denmark stops the onshore carbon storage project at the Vattenfall Nordjylland Coal Power Station.
127 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
in 2050. CCTS deployment starts once the CO2 price exceeds 40 €/tCO2 which happens in the year 2030 due to the steep path increase. The first investments into the CCTS technology are seen in the previous years (2020-2025). The iron and steel sector is – similar to modeling runs in Chapter 5 – again the first mover until some cement works start capturing CO2 from 2035 onwards (see Figure
6.3). At that point a certificate price of 75 €/tCO2 is reached and a total of 300 MtCO2 is annually stored in offshore hydrocarbon fields and saline aquifers. CCTS becomes economical for power plants and refineries as soon as the price exceeds 100 €/tCO2 in the year 2040. Still rising prices above 180
€/tCO2 in 2045 lead to additional economic incentives for more distant power plants to invest in further
CCTS deployment. Annual captured emissions sum up to more than 1 billion t CO2 from 2040 to 2045. These emissions are then transported via a pipeline network of 44,800 km to different storage locations. Total captured emissions start decreasing after 2045 due to the phase-out of several older power plants. 12.2 GtCO2 is stored in offshore storage sites until 2050. 55% of these emissions originate from industrial sources.
Figure 6.3: Captured CO2 emissions by source and storage type over time in the EU_80% scenario. The capturing costs have the highest share (60-70%) in variable as well as fixed costs of the CCTS chain. The infrastructure costs of storage comprise around 30% of the total investment costs, but have relatively small share of total variable costs of 10%. Transport costs depend very much on the location of each facility and range around 10-15% in variable and fixed costs. This step of the CCTS technology chain is also the driver making CCTS a more beneficial option for facilities closer to potential storage sites. This can be clearly seen as the first movers are mostly located near the North Sea where the highest offshore storage potential can be found. The overall investment costs until 2050 exceed €300 bn with an additional €730 bn of variable costs.
128 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
6.3.3 Sensitivity to investment and variable costs Many cost studies of the CCTS technology chain name the capture stage as most cost intensive for both investment and variable costs (e.g. The Crown Estate, Carbon Capture & Storage Association, and DECC 2013). To assess the sensitivity of the resulting CCTS infrastructure to these cost parameters, we simulated four additional scenarios: Two where we double the capital costs (Inv_200%) and variable costs (Var_200%) respectively, one with double capital and variable costs (Inv&Var_200%), and one with variable and capital costs both increased by 50% (Inv&Var_150%).99 Table 6.3 provides the input values for the sensitivity analysis and reference values from CCTS-Mod and the PRIMES model of the European Commission (EC 2013b) for comparison. The capital costs used for the base run are 25-30% below the input values in the PRIMES model. For variable costs no values for comparison were available. The results are calculated using a discount rate of 5%. In general, higher discount rates give lower weight to costs that occur further in the future. Investment cost are not annualized in CCTS-Mod but are incurred as a total five year before the infrastructure can be used. By contrast, variable costs occur on an annual basis. Therefore their relative importance investment cost increases with higher discount rates. Additional sensitivity analysis with 2.5% and 7.5% discount rate, however, confirm the relative importance of variable versus investment costs in all runs even though absolute importance diminishes with higher discount rates and increases with lower discount rates.
Table 6.3: Input parameters for sensitivity analysis, and reference values for comparison. Input Parameter Variation 2015 2020 2020 2030 2035 2040 2045 2050 Capital cost Base Case100 175 175 162 149 138 127 118 108 in €/tCO2y Inv&Var_150% 263 263 243 224 207 191 177 162 Inv_200% 350 350 324 298 276 254 236 216 PRIMES101 211 202 153 Variable cost Base Case100 64 64 63 62 61 60 59 58 in €/tCO2 Inv&Var_150% 96 96 95 93 92 90 89 87 Var_200% 128 128 126 124 122 120 118 116 Source: EC (2013b) and Mendelevitch (2014). In all sensitivity runs the increase in costs leads to a significant delay in initial deployment of CCTS technology. Figure 6.4 (left side) shows that while in the base run CCTS is first introduced in 2030, in the Inv_200% and Inv&Var_150% scenario the technology is first used in 2035, and only in 2040 in the other two scenarios. The figure also illustrates the sensitivity of total costs and length of the pipeline network until 2050. For all sensitivity runs cost figures are 5-25% higher than in the base case, showing an increasing sensitivity over the model horizon due to the accumulation of higher variable costs. Figures on CO2 capture, storage and pipeline network are lower for the sensitivity runs than for the base case, with the gap narrowing between 2040 and 2050 (see Figure 6.4 right side). For
99 The given costs only include the additional variable and fixed costs for equipping a power plant or industrial facility with a capturing unit compared to a facility without a capturing unit. 100 Data specification used for coal-fired power plants in Mendelevitch (2014). 101 EC (2013); based on emission factor 0.9 tCO2/MWh, load factor 86%, reference power plant 2100€/kW overnight capital costs.
129 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
the two scenarios Inv_200% and Inv&Var_150% the overall impact on key results like capture, and storage amounts and length of pipeline infrastructure is at most 10% or less. By contrast, doubling the variable capture costs has a strong impact on the length of the pipeline network with a decrease of over 35% compared to the base case. The future development of a CCTS infrastructure is therefore more sensitive to its variable costs than its investment costs. However, the deployment of the CCTS technology as a whole is not very sensitive to even drastic increases in capture costs, given high CO2 certificate prices in the end of the modeling horizon (270 €/tCO2) and the lack of alternative technologies, as both prevailing in this modeling framework.
Figure 6.4: Sensitivity of captured amounts over the model horizon (left side), and total costs and length of the pipeline network in 2050 (right side).
6.3.4 Summary of the European-wide scenarios Table 6.4 summarizes the results of the different Europe-wide scenarios. A summary of all scenario results can be found in Table B.3 in Appendix B. In the EU_40% scenario only four iron and steel factories in Norway and Estonia invest in the capture technology as they profit from the industry’s low variable and fixed costs. These facilities additionally benefit from their ideal location close to storage sites in the North Sea, minimizing costs associated with CO2 transport. CCTS cannot be considered as an abatement option for power plants if CO2 prices barely rise above 50 €/tCO2. Sensitivity analysis shows that the future development of a CCTS infrastructure is more sensitive to its variable costs than its investment costs.
The EU_80% scenario arrives at CO2 certificate prices around 250 €/tCO2 in the year 2050. Under this assumption, investing in the CCTS technology is cost-efficient for all emitters, with industry still being the first mover. However, from today’s perspective, these modeling results seem unrealistic. Even under the assumption of one omniscient planner, a CO2 pipeline network of at least 45,000 km covering great parts of Europe would be needed. Overall system costs, including costs of carbon capture, transport and storage over the entire model horizon, sum up to €800-1,000 bn. The construction of such a huge new infrastructure network is highly dependent on public acceptance, Keefe, and Mander 2014). Considering׳especially in densely populated regions like Europe (Gough, O the number of different parties, technology stages, insecurities regarding CO2 prices, learning rates and further policy measures, one comes to the conclusion that the necessary infrastructure and investment costs would be several times higher. This questions the fact whether CCTS may be able to
130 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
fulfill its role as a decarbonization technology within Europe. The following section 6.4 therefore focuses on regional CCTS deployment around the North Sea only.
Table 6.4: Summary of the European-wide results. Scenario Pipeline Number of Stored Origin. Storage CCTS CCTS network wells* emiss. until from left in invest. var. [1000s km] [GtCO2] industry 2050 costs costs 2030 2050 2030 2050 2030 2050 [%] [GtCO2] [€bn] [€bn] EU_40% - <1 0 2 - 0.02 100 50.0 0.2 0.4 EU_80% 1.4 45.3 34 1176 - 12.2 55 37.9 306.6 731.2
* assuming an annual well capacity of 0.8 MtCO2 following (IEAGHG and ZEP 2011).
6.4 Regional focus: CO2-enhanced oil recovery options in the North Sea and the Role of regional Cooperation The planned demonstration projects with the highest chance of realization are all close to the North
Sea and aim for offshore storage with additional profit generated from CO2-EOR (GCCSI 2014). The following scenarios depicted in sections 6.4.4 and 6.4.5 assess the implications of CO2-EOR for the development of a CCTS infrastructure with a focus on the North Sea Region. Several of these countries however, such as Germany and France, are unlikely to take part in any future CCTS deployment.102 Different national strategies towards implementation of CCTS, instead of a joint European energy strategy, thus seem most likely at the moment. Section 6.4.6 therefore includes a regionally focused analysis of four European countries where a joint CCTS and CO2-EOR deployment is most likely: Denmark, the Netherlands, Norway, and the UK (DNNU). One interesting aspect analyzed in this section is whether the employment of CO2-EOR by a limited number of countries increases costs due to a lack of economies of scale during the use of CO2-EOR and later. The assumed price paths are the same as in the previous scenarios.
6.4.1 The role of CO2 reuse for CCTS
CO2-EOR is the most mature CO2 reuse technology and has been practiced since the 1980s in the USA and Canada (cf. GCCSI 2011). The application of other technologies that are in the commercialization phase like Bauxite residue carbonation and using CO2 in methanol production is very site specific and requires favorable local conditions. The use of CO2 in enhanced coal bed methane recovery, as a working fluid in enhanced geothermal systems, as feedstock in polymer processing, and for algae cultivation are all technologies that need to be further developed and proven in real world pilot or demonstration scale applications. The global market for CO2 reuse across all technologies has a volume of approximately 80 Mt per year, which is equivalent to the annual emissions of the four biggest lignite power plants in Germany. CO2-EOR in the USA and Canada account for the biggest share with 50 Mt per year. 80% of the CO2 is supplied from natural CO2 sources at a price in the order of 15-19 US$/tCO2. In total, anthropogenic CO2 emissions can only be offset to a few percent from current and potential future demand for CO2 reuse. Although reuse has
102 This is partly due to rising public opposition (not in my backyard effect - NIMBY) as well as different national interests (e.g. France focusing on nuclear energy, Germany on the other hand on renewable energy sources).
131 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
very limited potential it can generate modest revenues for a selection of near term CCTS projects. Its impact to global CO2 abatement, however, depends on the application as e.g. CO2-EOR and using
CO2 in methanol production have no positive climate effect due to the latter burning of the product (Gale et al. 2015b).
IEA and UNIDO (2011) give a similar assessment of the role of CO2-EOR for the development of the CCTS technology appraising it as an important way to add value to a CCTS operation. The IEA
(2012a) is analyzing the role of this technology. It acknowledges that CO2-EOR not only offers a way to partly offset the costs of demonstrating CO2 capture but also to drive the evolution of CO2 transportation infrastructure, and incorporates opportunities for learning about certain aspects of CO2 storage in some regions. Several studies have looked into the economics of CO2-EOR on a regional and national scale: e.g. the application of the technology in the UK Central North Sea/Outer Moray Firth region (Kemp and Kasim 2013; Scottish Centre for Carbon Storage 2009) and the Norwegian continental shelf (Klokk et al. 2010), and have found substantial potential for the combination of the two technologies and associated benefits.
6.4.2 CO2-EOR resources in the North Sea
The analysis of the role of CO2-EOR for the development of a CCTS infrastructure requires a comprehensive estimation of the potential for CO2-EOR in the North Sea region. Mendelevitch (2014) performed an intensive literature review and presents own estimates to compile a consistent database of CO2-EOR potentials in the North Sea region. Data availability diverges significantly between the various countries of the North Sea Region. Therefore, different approaches have been chosen for each country. CO2 injection potentials are considered as the net amount of CO2 that can be stored during the CO2-EOR process and includes a constant recycling ratio of 40% following Gozalpour et al. (2005).
For the UK Mendelevitch (2014) finds 54 candidate fields with an estimated net injection potential ranging between 2 and 89 MtCO2 (Forties field). Total UK potential sums up to 572 MtCO2 which corresponds to 1733 Mbbl additional oil recovery potential. For Norway he identifies seven candidate fields with an estimated net injection potential ranging between 4 and 130 MtCO2 (Ekofisk field). Total storage potential in Norwegian oil fields in the North Sea add up to 314 MtCO2 which corresponds to an additional oil recovery potential of 951 Mbbl. For Denmark the study finds 14 candidate fields with an estimated net injection potential ranging between 3 and 88 MtCO2 (Dan field). Total storage potential in Danish oil fields sums up to 348 MtCO2, which corresponds to an additional oil recovery potential of 1054 Mbbl. Other riparian countries of the North Sea do not exhibit substantial oil resources and are therefore not included in the analysis.
6.4.3 Costs and revenue of CO2-EOR
To assess the economics of a prospective CO2-EOR infrastructure correctly, it is crucial to accurately estimate the costs associated with it. Mendelevitch (2014) draws on various case studies on CO2-EOR projects in the North Sea to develop an inventory of the main investment and operating costs components (see Table 6.5).
132 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Based on the cost components mentioned above investment costs add up to 103.9 €/tCO2 stored per year and operating costs add up to 36.8 €/tCO2 stored. Without costs for CO2 import the costs for oil supply from CO2-EOR in the North Sea Region are in the range of €12-17 per bbl incremental oil
(depending on site specific CO2 utilization rates), which is consistent with estimates from OECD and
IEA (2008b), giving a range of US$40-80 per bbl (including costs of CO2 supply) for long-term oil supply from CO2-EOR.
Expectations about the development of the crude oil price determine the attractiveness of CO2-EOR operations. The price not only has to cover investment and variable costs of incremental oil production but also has to refinance the capture and the transport of the CO2. DOE/IEA (2012) present a compilation of different oil price projections for the Western Texas Intermediate (WTI) crude oil price for the period up to 2035. The chosen medium oil price path represents an average of the price projections while the lower price path marks their lower bound. To provide a rough estimate of the profitability of combining CCTS with CO2-EOR, Table 6.6 compares cost and revenue items for a generic example. The sales price of additionally produced crude oil and the assumed CO2 certificate price (as negative opportunity costs) of the respective year constitute the potential revenue side. On the costs side, investment and variable costs for each of the stages of the CCTS technology chain are included. Even for the high first-of-a-kind investment costs assumed for 2015 and 2020 the combination of the two technologies yields considerable profit of 100 €/tCO2 and higher. The two most critical assumptions are the “bbl crude oil per tCO2 injected” conversion rate and assumptions on the future development of oil prices. Until now, CO2-EOR operations have only been performed onshore. Employing the technology in the North Sea is associated with additional technological and therefore also financial risk which is not taken into account in this calculation. 103
Table 6.5: CAPEX and OPEX cost components for CO2-EOR installation. CAPEX cost component € mn 1) Survey costs to examine the reservoir characteristics with respect to CO2- 1.50 EOR 2) Platform construction/restructuring costs to adapt to CO2-EOR requirements, including a) surface facilities costs to pretreat the CO2 before injection 17.5 b) recycle installments to separate, compress and re-inject CO2 7.1 3) Well drilling costs for new injection wells 52.5 4) Monitoring and verification facility 3% of CAPEX OPEX cost component € mn/MtCO2 1) Facility operation 5% of CAPEX 2) Oil production 12.1 3) CO2 recycling 5.2 4) CO2 compression and injection 8.7 5) Monitoring and verification 0.4 Source: Mendelevitch (2014).
103 A CO2 utilization rate of 0.33 tCO2/bbl (Mendelevitch 2014) and 1.25$/€ is being used. Additional capture costs for a coal-fired power plant equipped with post-combustion capture are calculated including a 5% discount rate and 30 years of operating life. The transport costs are estimated by assuming a 500 km long pipeline, with a lifetime of 30 years and a 5% discount rate. CO2-EOR equipment is expected to have a much shorter operating life of 10 years and the same discount rate of 5%.
133 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Table 6.6: Cost and revenue items for the deployment of combined CCTS and CO2-EOR technology. Input Parameter Variation 2015 2020 2030 2040 2050 Crude Oil Price in $/bbl 92 106 118 123 135 €/tCO2 222 255 282 294 324 Certificate price (40% €/tCO 14 17 37 52 52 Scenarios) 2 Subtotal: returns €/tCO2 236 272 319 346 376 Capture Capital cost €/tCO2y 175 175 149 127 108 Variable cost €/tCO2 64 64 62 60 58 Σ €/tCO2 75 75 72 68 65 Trans- Capital cost €/tCO2y 57 57 57 57 57 port Variable cost €/tCO2 5 5 5 5 5 Σ €/tCO2 9 9 9 9 9 Storage Capital cost €/tCO2y 104 104 104 104 104 Variable 104 €/tCO 37 37 37 37 37 cost 2 Σ €/tCO2 50 50 50 50 50 Subtotal: CCTS costs 134 134 131 127 124 Total: Returns – Costs €/tCO2 102 138 188 219 252 Source: Mendelevitch (2014).
6.4.4 Regional scenario: NorthSea_40% scenario with CO2-EOR option
The NorthSea_40% scenario assumes the same CO2 price path as the EU_40% scenario (see Table 6.1). Scenario results show that the use of CCTS is still most economical for the industrial sector, particularly iron and steel making plants. These facilities invest in a CCTS infrastructure from 2015 to
2020 in order to gain profits from additionally recovered oil from CO2-EOR from 2025 onward. After the exhaustion of most of the CO2-EOR fields in 2035, new storage sites in saline aquifers and depleted hydrocarbon fields closer to the shore are being used (see Figure 6.5 for the CO2 flows in 2050). In this scenario, a total of 2.5 bn tCO2 is stored until 2050 with annual storage volumes of around 100
MtCO2. The required CO2 transport network spans approximately 15,000 km. The scenario indicates that the CO2-EOR technology could lead to additional early economic incentives for the construction of a CCTS infrastructure. Existing infrastructure can be used after the exploitation of the CO2-EOR potential in the North Sea as soon as the CO2 price is high enough. In the case of CO2 prices remaining at around 50 €/tCO2 (as seen in the EU_40% scenario), it is still only economical for several industrial facilities such as steel or cement. The investment costs sum up to €50 bn with additional variable costs of €150 bn until 2050. Revenue from selling additionally recovered crude oil sums up to
€300 bn. Thus, even if investments in CCTS infrastructure are more than recovered, CO2 price signals above 50 €/tCO2 are needed in order to establish long-term use of CCTS.
104 Variable costs of CO2 storage include operational costs (OPEX) of oil production (see Table 6.5).
134 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Figure 6.5: CO2 flows in the NorthSea_40% scenario in 2050 after CO2-EOR-fields are exploited.
6.4.5 Regional scenario: NorthSea_80% scenario with CO2-EOR option
The NorthSea_80% scenario assumes the same CO2 price path as in the EU_80% scenario (see
Table 6.1). Until 2035 – the point when the CO2-EOR operation stops due to depletion – results of the NorthSea_80% scenario are very similar to those of the respective NorthSea_40% scenario. From
2020 onwards an average of 100 MtCO2 is transported each year from steel and cement facilities into
CO2-EOR operations in the North Sea. Once CO2-EOR resources are depleted, further CO2 capture activity is solely incentivized by the CO2 certificate price, which has to cover at least the variable costs as well as potential new investment costs. New storage in non-CO2-EOR locations is being developed close to the shore and close to already existing transport routes. From 2035 onwards, with prices exceeding 75 €/tCO2, additional more distant industrial facilities start running their capturing units.
Similar to the results from the respective EU_80% scenario without the CO2-EOR option, power plants only start capturing their CO2 from 2040 onward. The network required to accomplish the CO2 transport spans 27,000 km connecting the sources to the North Sea storage sites (see Figure 6.6). The investment costs sum up to €190 bn and there are an additional €540 bn variable costs over the whole time period until 2050 (see Figure 6.7). Revenues from selling additionally recovered crude oil sum up to €300 bn, similar to the results in the NorthSea_40% scenario. However, in contrast to the
NorthSea_40% scenario, the high CO2 price in this scenario creates enough incentive to pursue CCTS even after the depletion of CO2-EOR resources and eventually leads to full deployment of the technology in the modeled sectors.
135 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Figure 6.6: CO2 flows in the NorthSea_80% scenario in the year 2050 after CO2-EOR fields are exploited.
Figure 6.7: Cost distribution over the whole timespan in the NorthSea_80% scenario in €bn.
Note that the total amount of CO2 captured is lower than in the EU_80% scenarios without the CO2- EOR option because this analysis focuses only on the riparian countries of the North Sea. However, like in the EU_80% scenario, all examined industrial facilities and power plants start using the CCTS
136 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
technology at some time; with industry still holding the higher share of total stored emissions over time.
6.4.6 Regional scenario: DNNU_80% scenario focusing on CO2-EOR in DK, NL, NO and UK Against the background of negative public opinion towards CCTS and lack of industry and policy commitment in Germany, France, Belgium and Sweden, we examine an additional scenario where only Denmark, the Netherlands, Norway and the UK have the possibility to use the CCTS technology.
In contrast to the other European countries, these four have a higher potential to use captured CO2 to generate additional revenue in the domestic oil industry, or at least support the application of CCTS in the industry sector (like in the Netherlands). Moreover, UK and Norway are still the only two signatories to the amended London Protocol to allow transnational CO2 transport for offshore storage
(GCCSI 2014), and these four are among the most advanced countries ready for large-scale CO2 storage operation (GCCSI 2015a). Our goal is to compare these results to the results of the other scenarios and to examine to what extent CCTS deployment is reduced due to a lack of economies of scale.
Similar to the previous scenarios, the use of CCTS is mainly economical for the industrial sector, particularly iron and steel making plants. In the DNNU_80% scenario, facilities invest in a CCTS infrastructure from 2015 to 2020 in order to gain profits from additionally recovered oil from CO2-EOR from 2025 onward. Around 100 MtCO2 is stored annually until the full exhaustion of the CO2-EOR resources, 10 to 15 years after the beginning of the operation (with a concentration in the first ten years). From 2035 onwards, additional storage sites in saline aquifers and depleted hydrocarbon fields closer to the shore are used by industrial facilities already equipped with CO2 capture. With CO2 prices exceeding 75 €/tCO2 in the DNNU_80% scenario, additional, more distant industrial facilities start investing in capture units. Power plants only start using the CCTS chain from 2040 onwards, similar to the outcome of previous scenarios without the CO2-EOR option.
For the period of the CO2-EOR boom (2025-2035), the results of the DNNU_80% scenario on length of the pipeline network are similar to those of the NorthSea scenarios. While distances to deliver CO2 up to the shore are shorter on average, CO2 from the UK takes especially long routes offshore to arrive at CO2 storage sites with CO2-EOR options (see Figure 6.8). The overall installed pipeline network in 2030 covers over 11,000 km (10,200 in the NorthSea scenarios) Similarly, the values for average investment in CO2 transport and CO2 storage per MtCO2 per year during the initial phase in 2025 do not change for the DNNU scenario (cf. Table 6.7).105 Due to a similar deployment of the technology, no economies-of-scale effect between the NorthSea_80% scenario in 4.5 and the DNNU_80% scenario can be observed during this period. However, the DNNU_80% scenario exhibits a shift in CO2-EOR utilization. We find that UK CO2-EOR storage potential used by France and
Belgium in the other scenarios is now intensively used to store domestic CO2 from UK (increase of 46
105 To assess economies of scale for the CO2-EOR boom period one has to compare 2025 values from Table 6.7. Values for 2030 also include investments for non-CO2-EOR induced CO2 transport and storage, as investments the model features a 5 year construction period before infrastructure can be used.
137 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
MtCO2 per year for the period from 2025 to 2040 in the UK). The same effect but to a smaller extent (9
MtCO2 per year) can be observed in Norway. Danish oilfields that stored CO2 from Germany in the other scenarios, now increasingly receive CO2 from the Netherlands (increase of 27 MtCO2 captured per year in the Netherlands in the period from 2025 to 2040). At the same time, capture activity in
Denmark does not change significantly. After the CO2-EOR boom, the storage volumes for the four countries do not differ between the different scenarios. A clear economies-of-scale effect can be observed for the post-CO2-EOR period. In 2040, average investment costs in both CO2 transport and storage infrastructure are much higher for the DNNU scenario compared to the NorthSea scenarios.
CO2 storage costs increase by more than 30% in 2040 while transport costs even double (cf. Table 6.7). The constructed transport network is much smaller than in the NorthSea_80% scenario (13,600 km compared to 26,800 km) which is due to a smaller observed area and the lack of economies of scale. The Table 6.8 summarizes the key results of the NorthSea and DNNU scenarios. A summary of all scenario results can be found Table B.3 in the Appendix. Due to their regional focus, volumes of
CO2 stored and required transportation distances in these scenarios are likely to be shorter than in the European-wide scenarios of section 6.3.
Figure 6.8: CO2 flows in the DNNU_80% scenario in 2025 using the CO2-EOR-option (left) and in 2050 after CO2-EOR-fields are exploited (right).
Table 6.7: Average investment costs in CO2 transport and CO2 storage per MtCO2 per year, comparing the NorthSea_80% and DNNU_80% scenarios. Coverage 2025 2030 2035 2040
Avg. Invest. in CO2 Transport per All North Sea Region 0.07 0.09 0.11 0.03 MtCO2 per year DK, NL NO, UK 0.07 0.07 0.09 0.07 Avg. Invest. in CO2 Storage per All North Sea Region 0.10 0.11 0.16 0.10 MtCO2 per year DK, NL NO, UK 0.10 0.10 0.16 0.15
138 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
Table 6.8: Summary of regional results. Scenario Pipeline Number of Stored Origin. Storage CCTS CCTS Network wells* Emiss. until From left in invest. var. [1000s km] [GtCO2] industry 2050 Costs costs 2030 2050 2030 2050 2030 2050 [%] [GtCO2] [€bn] [€bn] NorthSea 14.2 15.4 119 141 0.6 2.5 100 40.0 47.2 150.0 _40% NorthSea 10.2 26.8 122 760 0.6 8.5 54 34.6 191.9 539.3 _80% DNNU 11.0 13.6 110 174 0.6 3.1 57 36.4 61.7 232.4 _80%
* annual well capacity for CO2-EOR: 1MtCO2; DOGF:0.8 MtCO2 ; Saline Aquifers: 0.8 MtCO2 following (IEAGHG and ZEP 2011) and own assumptions.
6.5 Conclusion: the importance of CO2-EOR for a European CCTS roll-out This chapter presents scenario analyses and interpretation on the potential role of CCTS to support the EU energy system transition to meet emission reductions goals that are consistent with the international goal of staying below 2°C of global warming. The assumptions of the moderate scenarios include a CO2 price of 50 €/tCO2 in 2050 which triggers hardly any CCTS development in Europe.
Additional revenues from applying CO2 enhanced oil recovery (CO2-EOR) in the North Sea lead to an earlier adoption of CCTS starting in 2025 independent from the CO2 certificate price. The lifespan of most CO2-EOR operations is expected to be around ten years. It is followed by conventional CO2 storage in nearby depleted hydrocarbon fields and saline aquifers if the CO2 certificate price exceeds the sector-specific thresholds to cover variable costs of carbon capture. Generally, the use of CO2 for
EOR projects is criticized by environmental organizations, as the overall CO2 mitigation effect is negative considering the CO2 content of the additional extracted oil.
More stringent climate scenarios aim at an 80% GHG reduction until 2050. The resulting CO2 price of
270 €/tCO2 in 2050 pushes all EU-ETS industry and energy sectors to use CCTS at some point. It is, however, the iron and steel sector that start deployment as soon as the CO2 certificate price rises above 50 €/tCO2 in 2030. The cement sector follows some years later at a threshold of around 75
€/tCO2. It is only with CO2 certificate prices exceeding 100 €/tCO2 that CCTS becomes lucrative for the electricity sector. Sensitivity analysis shows that the future development of a CCTS infrastructure is more sensitive to its variable costs than its investment costs. As European usage of onshore storage sites is unlikely due to high public resistance, transport distances increase. The resulting CO2 transport network required to connect emission sources and storage sites across Europe would comprise of up to 45,000 km of pipeline and store up to 1,000 MtCO2 per year.
Taking into account the realities that confront CCTS in the EU, political and public opposition has left only a handful of countries that still consider building CCTS in the medium term. A 20% CCTS penetration rate in the European power sector as calculated in the DNNU_80% scenario in 2050 thus seems more realistic. Concentrating on Denmark, the Netherlands, Norway and the UK, this scenario shows an increased utilization of CCTS-EOR especially in the UK and the Netherlands. However, a
139 Chapter 6: European Scenarios of CO2 Infrastructure Investment until 2050
lack of economies of scale leads to increasing average costs, once the CO2-EOR-fields have been exploited: CO2 storage costs increase by more than 30% in 2040 while transport costs even double.
A critical point of our analysis is that the employed model CCTS-Mod is purely cost-driven and does not include any specific bound on the CCTS penetration. The model assumes a cost-minimizing player with perfect foresight and therefore tends to overestimate the potential for CCTS. Additional legal, political, and environmental issues with respect to transboundary CO2 transport as well as CO2 storage and liability issues are not included in the model. The model analysis highlights that the international distribution of CO2-EOR and non-EOR storage sites leaves room for significant international conflicts of interest and the need for coordination between the North Sea riparian countries. Some countries, e.g. France or Germany, only have limited (offshore) CO2-storage capacities. A non-coordinated European CCTS utilization does not profit from the mentioned economies of scale resulting in higher overall system costs. Real costs are expected to be higher and come with a lower deployment of CCTS in the future. On the other hand, already the existence of a
CO2 infrastructure might impact investment decisions in various sectors of the energy system (spatially and with respect to the selection of generation technology) that are neglected in this modeling approach. Considering the large number of different players and technologies, the insecurities regarding CO2 prices, learning rates, legal issues, public resistance and further policy measures strongly question whether CCTS may be able to fulfill its role as a major bridging technology for the decarbonization of Europe.
CO2-EOR is the driver behind all global CCTS projects that will become operational in the near future or have already started operation (e.g. Boundary Dam, Canada). The underlying regulatory frameworks and support schemes can primarily be regarded as a cross-subsidization of the petroleum industry, while progressing the CCTS technology is of secondary interest. This is underpinned by observations in the Gulf States, USA and Canada, where the legislative framework for CO2-EOR with
CO2 recycling is established, while the framework for long-term storage (which would be the primary goal of CCTS) is underdeveloped. The same is true for Europe, where the emergence of a regionally focused network around the North Sea, including only a few riparian countries using offshore CO2 storage with CO2-EOR, is the most likely option. The mirage of a Pan-European network for CCTS in the EU-ETS industry and energy sectors, as envisioned in some long-term scenario projections, seems out of reach at present due to a combination of a lack of financial incentives as well as too little political and public support for CCTS as a mitigation technology. Further research, however, is needed to evaluate the effects of the newest European reforms (e.g. the reform of the EU Emissions Trading System ETS) as well as national regulations (e.g. emissions performance standards (EPS) and contract for differences (CfD) in the UK) on the development of CCTS.
140 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
Chapter 7 THE IMPACT OF POLICY MEASURES ON FUTURE POWER GENERATION PORTFOLIO AND INFRASTRUCTURE – A COMBINED ELECTRICITY AND CCTS INVESTMENT AND DISPATCH MODEL (ELCO)
7.1 Introduction: a review of state of the art electricity and CO2 modeling approaches The need for combating climate change is internationally widely accepted (World Summit of the Regions 2014) and the role of the electricity sector as a major contributor to global GHG emission reductions is undisputed (Leader of the G7 2015).106 However, there exists an international dissent on how to achieve a decarbonization of the sector. Even in the EU, a multitude of approaches exist: Germany has departed on its “Energiewende” path towards a renewable energy based system, with renewable energy sources (RES) already contributing to 30% of electricity production in 2015. At the same time, France still relies on large nuclear capacities; while the United Kingdom (UK) promotes a mixed strategy of renewables, nuclear and carbon capture, transport, and storage (CCTS). The low certificate prices in the European Emissions Trading System (EU-ETS), at levels below 10 €/tCO2 in 2015 – with little hope for a significant rise in the upcoming years (Hu et al. 2015) – however, give insufficient incentives for most of these low-carbon investments. This endangers achieving the EU climate policy targets for 2030 (EC 2014b) and puts the global 2°C target at risk. Therefore, several countries have started or are about to start backing the EU-ETS with additional national measures.
These include different types of feed-in tariffs and market premia, capacity markets, a minimum CO2
106 This chapter is based on an article in the IEEE Conference Publications for the 12th International Conference on the European Energy Market (EEM), Lisbon, Portugal, 2015 (Mendelevitch and Oei 2015). It is joint work with Pao-Yu Oei and was started during a research stay at the International Institute for Applied Systems Analysis (IIASA) in Laxenburg, Austria in the autumn of 2014. Roman Mendelevitch and Pao-Yu Oei jointly developed the model and its implementation in GAMS. Pao-Yu Oei was in charge of the implementation of the UK case study. Roman Mendelevitch had the lead in collecting data. The writing of the manuscript was executed jointly.
141 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
price and emissions performance standards (EPS). Models assessing the future development of a decarbonized electricity market need to adequately incorporate such additional policy measures. In addition, interdependencies between the measures as well as feedbacks with other sectors need to be taken into account.
Different kinds of models are used to assess the impact of policy instruments and their ability to achieve climate change policy objectives. Pfenninger et al. (2014) classify models according to the different challenges they address. They differentiate between energy system models for normative scenarios, energy system simulation models for forecasts, power systems and electricity market models for analyzing operational decisions and qualitative and mixed-methods for narrative scenarios. Energy system models such as PRIMES (Capros et al. 1998), MARKAL (Fishbone and Abilock 1981), EFOM (Finon 1979) or POLES (Criqui 1996) are able to convey the “big picture” of what is happening in different linked sectors of an energy system. These technology-oriented models focus on the energy conversion system, on the demand-side (e.g. efficiency measures) as well as supply side (e.g. wide range of generation technologies). The advantages of these models are that they cover several sectors, linking them through endogenous fuel substitution. They are mostly solved by optimization or simulation techniques when minimizing system costs or maximizing the overall welfare. Fais et al. (2014) integrate different types of RES support schemes such as feed-in tariffs as well as quantity based instruments such as certificate systems in their energy system model Times-D. Their approach can be used to analyze exogenous support scheme but does not establish a link between attaining a specific CO2 target and the level of required RES support, and does not allow analysis of long-term development. Moreover, RES generation is limited exogenously via upper bounds on annual maximum expansion. They assume perfect competition and have limited possibilities to incorporate market power.
Apart from energy system models, there is a large strand of literature that employs a partial equilibrium setting to assess one particular market, e.g. the electricity market. This allows for analyzing non- cooperative firm behavior in more detail (e.g. à la Cournot) by allowing the firms to strategically exploit their influence on the market price with their output decision. Moreover, different risk attitudes and explicit shadow prices can be easily incorporated in these settings. The models have been focusing on considerations of resource adequacy (Ehrenmann and Smeers 2011), assessing the impact of environmental regulation (Allevi, Bonenti, and Oggioni 2013), renewables obligations and portfolio standards (see e.g. Gürkan and Langestraat 2014; Chen and Wang 2013), or congestion management of the transmission network (Kunz and Zerrahn 2015).
One technology that is of particular interest for a future decarbonization of the electricity sector is CCTS. The technology comes with a dichotomy: On the one hand, it plays an important role in many of the possible energy system scenarios that are consistent with the EU Energy Roadmap (EC 2013b). Accordingly, the scenarios for the newest report from the IPCC (2014) estimate a cost increase of 29-
142 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
297% for reaching the 2°C target without the CCTS technology.107 On the other hand, despite available financial schemes and technology, CCTS has not been implemented on a large scale anywhere in the world. Various authors have addressed this discrepancy with different regional focuses (Groenenberg and de Coninck 2008; Hirschhausen, Herold, and Oei 2012b; Milligan 2014; Stechow, Watson, and Praetorius 2011). Gale et al. (2015b) in addition address this topic in a special issue commemorating the 10th anniversary of the first IPCC (2005) special report on CCTS.
Most electricity market models do not put any emphasis on CCTS, and handle the technology like any other conventional generation technology by specifying investment and variable costs and fuel efficiency. For example, Eide et al. (2014) apply a stochastic generation expansion model to determine the impact of CO2 EPS on electricity generation investment decisions in the U.S. Their findings show a shift from fossil fuel generation from coal to natural gas rather than incentivizing investment in CCTS. Zhai and Rubin (2013) explored the “tipping point” in natural gas prices for which a coal plant with CCTS becomes economically competitive, as a function of an EPS. Middleton and Eccles (2013) calculate the price for CO2 to be in the range of 85-135 US$/tCO2 (65-105 €/tCO2) to incentivize a gas power plant to use CCTS in the USA. This simplified representation of the CCTS technology in these models, however, neglects transportation and storage aspects as well as the possibility of industrial usage of CCTS.
By contrast, if models focus on CCTS infrastructure development, they often neglect how the technology is driven by decisions in the electricity market. A series of studies analyzed the technical potential of CCTS deployment, including possible CO2 pipeline routing (Oei, Herold, and Mendelevitch 2014; Morbee, Serpa, and Tzimas 2012; R. S. Middleton and Bielicki 2009; Kazmierczak et al. 2008; Kobos et al. 2007). The construction of such large-scale new infrastructure networks is highly influenced by public acceptance, especially in densely populated regions such as the European Union Keefe, and Mander 2014). Acceptance issues as well as other technical uncertainties can׳Gough, O) lead to high cost increases of a CCTS deployment (Knoope, Ramírez, and Faaij 2015). In the absence of expected technological learning and with persistently low CO2 certificate prices CCTS projects aim at additional income through CO2-Enhanced Oil Recovery (CO2-EOR) (Mendelevitch 2014; Kemp and Kasim 2013).
Kjärstad et al. (2013) have started to close this gap by combining the techno-economic Chalmers
Electricity Investment Model with InfraCCS, a cost optimization tool for bulk CO2 pipelines along with
Chalmers databases on power plants and CO2 storage sites. Their approach, however, relies on solving both sectors consecutively starting with the electricity model without any feedback options.
They, in addition, do not include CO2 capture from industrial sources. This neglects economies of scale especially with respect to transporting CO2 as well as scarcity effects with respect to CO2
107 RES and nuclear provide sufficient decarbonization alternatives for the electricity sector. The high cost increase, however, is caused by only limited alternative decarbonization technologies in the industry sector. Negative emissions of large-scale utilization of CCTS with biomass, in addition, compensate for unabatable emissions in other sectors (Kemper 2015).
143 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
storage. Additional research is needed to include different policy instruments into the modeling frameworks to evaluate the effect of various measures.
This chapter presents a general electricity-CO2 (ELCO) modeling framework that is able to simulate interactions of the electricity-only market with different forms for national policy measures as well as a full representation of the carbon capture, transport, and storage (CCTS) chain. Different measures included in the model are feed-in tariffs, a minimum CO2 price and a CO2 emissions performance standard (EPS). Additionally, the model includes large industrial emitters from the iron/steel and cement sector that might also invest in carbon captures facilities, increasing scarcity effects for CO2 storage. The set-up also takes into account demand variation by type hours, the availability of more and less favorable locations for RES and endogenously accounts for limits to annual diffusion of new technologies. The model is driven by a CO2 target and an optional RES target. This chapter is used to describe the different features and potentials of the ELCO model. We apply the model to a stylized case study of the UK Electricity Market Reform (EMR) to present a show case of our model framework.
The remaining chapter is structured as follows: The introduction is followed by a detailed description of the ELCO model in section 7.2. A case study in section 7.3 applies the ELCO model to the UK electricity market. The main policy measures are adjusted in the model to mimic the UK EMR and its long-term effects. Section 7.4 concludes with an outlook of future applications of the ELCO model.
7.2 Mathematical representation of the ELCO model The ELCO model mimics the competition of different conventional electricity generation technologies on the electricity market and their interaction with new technologies that are financed via fixed tariffs. Each technology is represented via a stylized player that competes with one another. For a better representation of scarce CO2 storage resources we also include a detailed representation of the complete CCTS value chain. This also includes potential CO2 capture from the steel and cement industry. The different CO2 storage options such as CO2-EOR, saline aquifers and depleted oil and gas reservoirs compete against one another in the last stage of the CCTS value chain. All players maximize their respective profits subject to their own as well as joint technical and environmental constraints. Other (external) costs as well as further welfare components are not being analyzed. Regional disaggregation takes into account geographical characteristics like availability (especially with respect to maximum potential and conditions for renewables as well as CO2 storage) and specific electricity demand.
Different policy measures such as a Carbon Price Floor (CPF), an Emissions Performance Standard (EPS) or feed-in tariffs in form of Contracts for Differences (CfD) are included in the modeling framework. The ELCO model analyzes how these policy instruments will influence the construction of new generation capacities. CfD for newly constructed low-carbon technologies can be derived endogenously using shadow variables of constraints. Assuming perfect competition between the different players, equilibrium is reached when overall system costs are being minimized subject to all constraints.
144 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
The developed model is able to assess regionally disaggregated investment in electricity generation, generation dispatch and simplified flows as well as CO2 transport, storage, and usage for CO2-EOR.
Incorporating CO2 capture by industrial facilities from the steel, and cement sector enables, on the one hand, the representation of economies of scale along the transport routes while, on the other hand, leading to higher scarcity effects with respect to CO2 storage options.
7.2.1 Notations of the model The following tables list the used sets, variables and parameters of the ELCO Model. Parameters are indicated by capital letters, variables by small sized letters and sets are resembled in subscripts. The detailed Karush-Kuhn-Tucker (KKT) conditions of the ELCO model are depicted in Appendix B.2.
Table 7.1: List of sets of the ELCO model. Name Description a, aa, aaa 5 year period h, hh Time interval i, ii CO2 sources from industry {Steel: IND_ST, Cement: IND_CE} n, nn Node new(t) Flag if a technology is newly built {0,1} s, ss CO2 sinks {Saline: STO_SA, DOGF: STO_DA, EOR: STO_SA} Generation technologies: { - g-type existing capacities: Nuc, Coal, Gas_GT: CCGT, Gas_CC: OCGT; - g-type new capacities: COAL_NEW, CCGT_NEW, OCGT_NEW; - g_cfd-type new capacities: PV: RES_PV, Wind_on: RES_WI_ON, Wind_off: t, tt RES_WI_OF, Hydro: RES_HY, Biomass: RES_BI, Coal_CCTS, CCGT_CCTS}
Table 7.2: List of parameters of the ELCO model. Name Description Unit ADJ_CO2(n,nn) Flag if two CO2-nodes are adjacent {0,1} ADJ_EL(n,nn) Flag if two Elec-nodes are adjacent {0,1} ALPHA(t,a) Maximal marginal CO2-abatement [ktCO2/GWh] AVAIL(h,n,t) Availability of power plant [%] CO2_IND(h,n,i,a) CO2 emission by industry [ktCO2] CO2_TARGET(a) CO2 target reduction for electricity sources [%] CP_CO2(s/i) Planning and construction period [years] CP_G(t) Planning and construction period [years] CPS(a) Carbon price support [k€/ktCO2] CR_G(t) Capture rate for generation 90% or 0% CR_IND(i) Capture rate for industries 90% D(h,n,a) Electricity demand [GW] DF(a) Discount factor [%] DIFF_CO2(s/i) Technology diffusion factor storage / industry capture [%] DIFF_G(t) Technology diffusion factor by generation technology [%] EF_EL(t) Emissions factor [ktCO2/GWh] EFF_CO2 CO2-EOR efficiency [kbbl/ktCO2] EUA(a) EU-ETS allowances [k€/ktCO2] FC_CO2(n,s/i,a) Fix costs for CO2 capture, and storage [k€/ktCO2] FC_CO2_T(n,nn) Fix costs for CO2 transport [k€/ktCO2] FC_F_E(n,nn) Fix costs for electricity transport [k€/GW] FC_G(n,t,a) Fix costs for generation w/o. or w/ capture [k€/GW] Flag if capacity investment from year a can be used for I_USE_CO2(s/i,a,aa) generation in year aa in the CO2 sector {0,1} I_USE_EL(t,a,aa) Flag if capacity investment from year a can be used for {0,1}
145 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
Name Description Unit generation in year aa in the electricity sector INICAP_EL_T(n,nn) Initial capacity for electricity transport [GW] INICAP_G(n,t,a) Initial capacity incl. retirement [GW] 2 INTC_CO2(t) Quadratic cost term for CO2 operation [k€/GWh ] INTC_G(t) Quadratic integration costs for generation technologies [k€/GWh2] Investment cost for industrial CO2 capture capacity or INVC_CO2(n,s/i,a) storage per hour [k€/ktCO2/h] INVC_CO2_T(n,nn) Investment cost for CO2 transport [k€/ktCO2/h] INVC_EL_T(n,nn) Investment cost for electricity transport [k€/GW] Investment cost for generation capacity w/o or w/ INVC_G(n,t,a) capture [k€/GW] LT_CO2(s/i) Life time of industry CO2 capture & storage technology [years] LT_G(t) Life time of generation technology [years] MAX_INV(n,t) Maximal potential of generation technology [GW] MAX_STOR(n,s) Maximal CO2 storage capacity [ktCO2] OILPRICE(a) Price of additional oil from CO2-EOR [k€/kbbl] ONE_FUEL(t,tt) Flag for identical fuel {0,1} PD(a) Period duration (5 years) [years] RE_TARGET(a) Renewables target [%] REF_CO2 CO2 emissions from electricity generation in 1990 [ktCO2] RES_OLD(h,n,a) Generation of already existing RE [GW] SP(t,a) Strike price for CfD-technologies in first years [k€/GWh] START_CO2(s/i) Starting capacity industry capture & storage technology [ktCO2/h] START_G(t) Starting capacity for generation technology [GW] TD(h) Time duration of each hourly segment [hours] Flag if capacity investment from years aa can be used USE_CO2(s/i,a,aa) for generation in year a in the CO2 sector {0,1} Flag if capacity investment from years aa can be used USE_EL(t,a,aa) for generation in year a in the electricity sector {0,1} VC_CO2(n,s/i,a) Variable costs for CO2 capture or storage [k€/ktCO2] VC_CO2_T(n,nn) Variable costs for CO2 transport [k€/ktCO2] VC_EL_T(n,nn) Variable costs for electricity transport [k€/GW] VC_G(n,t,a) Variable generation costs w/o. or w/ capture [k€/GWh]
Table 7.3: List of variables of the ELCO Model Name Description Unit CO2_c(h,n,i,a) Emissions captured from industry [ktCO2/h] CO2_s(h,n,s,a) Storaged emissions [ktCO2/h] CO2_t(h,n,nn,a) Flow of CO2 [ktCO2] el_t(h,n,nn,a) Flow of electricity [GW] emps(a) Emissions Performance Standard [ktCO2/GWh] g(h,n,t,a) Generation of electricity [GW] g_cfd(h,n,t,aa,a) Generation electricity from CfD sources [GW] inv_CO2_c(n,i,a) Investment in capture technology [k€/ktCO2/h] inv_CO2_s(n,s,a) Investment in storage technology [k€/ktCO2/h] inv_CO2_t(n,nn,a) Investment in CO2 transport capacity [k€/ktCO2/h] inv_el_t(n,nn,a) Investment in electricity transport capacity [k€/GW] inv_g(n,t,a) Investment in generation capacity [k€/GW]
146 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
Table 7.4: List of dual variables of the ELCO Model Name Description Unit lambda_cap_CO2_c(h,n,i,a) Dual of CO2 capture cap. [k€/ktCO2/h] lambda_cap_CO2_s(h,n,s,a) Dual of CO2 annual storage cap. [k€/ktCO2/h] lambda_cap_CO2_t(h,n,nn,a) Dual of CO2 transport cap. [k€/ktCO2/h] lambda_cap_el_t(h,n,nn,a) Dual of transmission cap. [k€/GW] lambda_cap_g(h,n,t,a) Dual of elec. generation cap. [k€/GW] lambda_cap_g_cfd(h,n,t,aa,a) Dual of elec. must run condition for RES [k€/GW] lambda_curt_el(h,a) Dual of electricity curtailment [k€/GWh] lambda_diff_CO2_c(i,a) Dual of diffusion for CO2 capture in industry [k€/ktCO2/h] lambda_diff_CO2_s(s,a) Dual of diffusion for CO2 storage [k€/ktCO2/h] lambda_diff_g(t,a) Dual of diffusion for renewables [k€/GWh] lambda_emps(n,t,a) Dual of emps constraint [k€/ktCO2] lambda_max_ind(h,n,i,a) Dual of maximum industry emissions [k€/ktCO2/h] lambda_max_stor(n,s,a) Dual of max. CO2 storage cap. [k€/ktCO2/h] lambda_pot_g(n,t,a) Dual of potential for renewables [k€/GW] lambda_target_CO2(a) Dual of CO2 emissions constraint [k€/ktCO2] lambda_target_RE(a) Dual of renewables target constraint [k€/GWh] mu_CO2(h,n,a) Dual of CO2 market clearing [k€/ktCO2/h] mu_el(h,n,a) Dual of electricity market clearing [k€/GWh]
7.2.2 The electricity sector
mu_ ehna,, gh,,, n t a EF_ ELt CPS a EUA a VC__ G INTC G· g n,,,,, t a t h n t a · target_2 co t, aaa aaa aaa I__ USE ELt,, aa aaa target_ RE 1_TARGET REaaa· aaa aaa I__, USE EL t,, aa aaa TD · tT_ RES h h target_ RE TARGET_ REaaa· aaa g_ cfd aaa I__, USE ELt,, aa aaa ELCE h,,,, n t aa a t T_ RES DF PD aa USE_ EL g/ g_ cfd a a t,, a aa a SP t, aa EF_ ELt 1 CR _ G t CPS a EUA a EF_ ELt CR _ G t mu _ co 2 h,, n a VC_ G INTC_ G· g_ cfd n, t,,,,, a t h n t aa a FC____ G INICAP G FC G inv g nta,,,,,,,, nta ntaa ntaa aa USE_ ELt,, a aa INVC__ G inv g n,,,, t aa n t aa aa USE_ ELt,, a aa (7.1)
The ELCO model represents electricity generation from various technologies. Electricity generation is herby divided in the two subgroups gh,n,t,a and g_cfdh,n,t,aa,a. gh,n,t,a comprise generation from all existing capacities and newly built carbon-intensive capacities from coal, gas OCGT and gas CCGT. g_cfdh,n,t,aa,a, on the other hand, include generation from newly constructed low-carbon generation capacities from PV, wind on/offshore, hydropower, biomass, CCTS coal/gas, and nuclear that are financed via the CfD scheme. The profit function for different technologies share the common
147 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
component of fix costs FC_Gn,t,a and annualized investment costs INVC_Gn,t,a depending on the investments inv_gn,t,a (lowest rectangular segment). The variable costs components and revenue differ: for g-type technologies (upper rectangle with upper flat corners) revenue is generated from sales on the electricity market receiving the electricity price mu_eh,n,a. The variable cost function comprise fuel and O&M costs with a linear and a quadratic term (VC_Gn,t,a and INTC_Gt). In addition
CO2 costs are calculated based on the emission factor EF_ELt, multiplied with a combination of the
EU-ETS CO2 certificate price (EUAa) and a carbon price support (CPSa in case of a carbon floor price for the electricity sector). For g_cfd-type technologies (middle rectangle with rounded corners) revenue is generated from the new CfD scheme. The CfD strike price can be incorporated in two ways: It can either be set exogenously, differentiated by year of construction and technology type. Or the strike price is determined endogenously. In the latter case, it depends on the extent to which generation from the respective technology contributes to achieving the environmental goals (TARGET_CO2a and
TARGET_REa) and is incorporated in the dual variables of these constraints (see section 7.2.2.1).
This type also encounters additional variable cost components for possible CO2 infrastructure
(transport and storage) which are passed via the dual variable mu_CO2h,n,a and account for CO2 capture rates CR_Gt. The technology specific quadratic cost term is interpreted as integration cost for increasing shares of g_cfd-type generation.
0_AVAILh,,,, n t TDh· inv g n tt aa EMPS aa h aa USE_, ELt,, a aa (,)_t tt ONE FUELt, tt gh,,, n t a EF_ EL t 1 CR _ Gt (7.2) emps TDh· n,, t a 0 g_ cfdh,,,, n tt aa a EF _ EL tt 1 CR_ G tt h aa USE_, ELt,, a aa t,_ tt ONE FUELt, tt
The individual players maximize their profit subject to several constraints. The EPS constraint (7.2) ensures that newly constructed generation capacities do not exceed the annual allowed CO2 emissions per GW. The overall emissions are calculated as an annual fuel and site specific sum, allowing for combined accounting of new capacities with and without CCTS.
The generation capacity constraints (7.3) and (7.4) differ slightly for conventional generation technologies gh,n,t,a and newly constructed low-carbon technologies g_cfdh,n,t,aa,a, as the calculation of currently available generation capacity differs for the two cases.
0AVAIL INICAP _ G inv _ g g cap_ g 0 h,,,,,,,,,,, n t n ta ntaa hnta hnta, (7.3) aaUSE_ ELt,, a aa
cap__ g cfd 0AVAILhnt,,,,,,,,,,,, inv _ g ntaa g _ cfd hntaaa hntaaa 0 (7.4)
148 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
A diffusion constraint restricts the maximal annual investment depending on generation from previous periods and some initial starting value for new technologies.
AVAILh,, n t·TD h hn, 0START _ Gt ··· TDh g _ cfd hntaaa, , , , 1 g _ cfd hntaa , , ,,at2 DIFF_ G #of nodes h,, n aa (7.5) diff_ g TDh· g _cfdh,, n t,,aa a t,a 0 h, n,a a
Another constraint limits the overall investment depending on a technology-specific maximal potential for each node.
pot_ g 0MAX _ INVn,,,,, t inv _ g n t aa n t a 0 (7.6) aaUSE_ ELt,, a aa
7.2.2.1 Shared environmental constraints for the electricity sector
All players in the electricity sector have to respect shared environmental constraints: An annual CO2 target guarantees that the annual dispatch is lower or equal an exogenously set CO2 reduction path.
0PD· TD g g _ cfd target_2 co 0 (7.7) athhnta,,,,,,, hntaaa,a a h, n,_ t aa USE ELt,, a aa
ALPHAt,a corresponds to the marginal contribution of the respective technology to the targeted CO2 intensity for a particular year. It is positive for low-carbon technologies while having negative values for conventional generation.
CO2 _ TARGETa REF _ CO 2 t, a 1 CR _ G t EF _ ELt (7.8) Dhna,, TDh hn,
National renewable targets setting a minimum share of renewable generation are implemented in an additional renewable constraint in some scenarios. This constraint, however, is deactivated in the scenario analyzed in this chapter.
g__ cfdh,,,,,, n t aa a RES OLD h n a aa USE_, ELt,, a aa tT _ RES target_ RE 0PDah··TD a 0 (7.9) hn, RE_ TARGET d a h,, n a hn,
7.2.3 The electricity transportation utility The objective function of the electricity transportation utility is shown in the following equation: The sum of variable costs VC_EL_Tn,nn and annualized investment costs INVC_EL_Tn,nn equalize the hourly electricity price difference between two nodes in case of no line congestion. Possible congestion rents are kept by the transportation utility as profit. Electricity is treated as a normal transport commodity ignoring Kirchhoff`s 2nd law as network congestion is not the focus of the ELCO model.
149 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
mu___ ehna,,,,,,, mu e hnna el t hnnna TDh TSO_ E h VC___ EL Tn,,,, nn el t h n nn a (7.10) DFaa PD a n, nn INVC____ EL Tn,,, nn inv el t n nn a aa a
The electricity utility maximizes its profits subject to the following line capacity constraint:
0INICAP _ EL _ Tn,,,,,,, nn ADJ _ EL n nn inv _ el _ t n nn aa ADJ _ EL nn n inv _ el _ t nn n aa aa a el_ t cap__ el t 0 h,,,,,, n nn a h n nn a (7.11)
7.2.4 The industry sector The industry is being represented by the two sectors i: Iron and Steel as well as cement which are most likely to use CO2 capture as mitigation option. The objective function of the industry sectors is limited to the abatement costs linked to exogenously given historic CO2 emissions. They include the option of either paying the EUAa or investing into the CCTS technology with its variable costs
VC_CO2n,i,a, fix costs FC_CO2n,i,a and annualized investment costs INVC_CO2n,i,a. The additional costs for a possible CO2 infrastructure (transport and storage) are being passed on from the downstream
CO2 sector via the dual variable mu_CO2h,n,a.
TDh CO2 _ IND h,,,,,, n i a co 2 _ c h n i a EUA a co2 _ ch,,,,, n i a mu _ co 2 h n a h co2 _ c VC _ CO 2 h,,,,, n i a n i a IND DFa PDa (7.12) a FC_ CO 2 inv _ co 2 _ c n,,,, i a n i aa aa USE_2 CO i,, a aa INVC_ CO 2 inv _ co2_ c n,,ia n,, i a aa USE_2 CO i,, a aa
The industry sector maximizes its objective function subject to similar constraints as the electricity sector. A diffusion constraint restricts the maximal annual investment depending on previous investments.
diff_ co 2 _ c 0START _ CO 2i inv _ co 2 _ c niaa,,,,, DIFF_2 CO i inv _ co 2 _ c nia ia 0 (7.13) n aa a n
The annual capturing quantity is restricted by the amount of previous investments as well as the overall maximal capturing quantity per node and technology.
cap_ co 2 _ c 0 inv _ co 2n,,,,,, i aa CR _ IND i co 2 _ ch n i a h n,, i a 0 (7.14) aaUSE _2CO i,, a aa
max_ ind 0CO 2_ INDhnia,,,,,,,,, CR _ IND i co 2_ c hnia hnia 0 (7.15)
150 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
7.2.5 The CO2 transportation utility
The CO2 transportation utility maximizes its profit show in Equation (15). The sum of variable costs
VC_CO2_Tn,nn and annualized investment costs INVC_CO2n,nn equalize the difference between the dual prices between two nodes.
mu_ co 2hna,,,,,,, mu _ co 2 hnna co 2_ t hnnna TDh TSO_2 CO h VC_ CO 2_ Tn,,,, nn co 2_ t h n nn a (7.16) DFaa PD a n, nn INVC_ CO 2_ Tn,,, nn inv _ co 2_ t n nn a aa a
A pipeline capacity constraint restricts CO2 transport:
ADJ_ CO 2 inv _ co 2_ t 0INICAP _ CO 2_ T n,,, nn n nn aa co 2_ t n,,,, nn h n nn a aa a ADJ_ CO 2nn,,, n inv _ co 2_ t nn n aa (7.17) cap_ co 2 _ t h,,, n nn a 0
7.2.6 The storage sector
Saline aquifers, depleted oil and gas fields (DOGF) and fields with the opportunity for CO2-EOR are identified as possible storage locations s. The objective function of the storage operator represents the abatement costs linked to the underground storage of CO2. For CO2-EOR sites it includes the option of returns received from oil sales at oil price OILPRICEa. The storage costs consist of the variable costs
VC_CO2n,s,a, a quadratic cost term INTC_St, fix costs FC_CO2n,s,a and annualized investment costs
INVC_CO2n,s,a. The dual variable mu_CO2h,n,a is used to pass on the overall storage costs (or in case of CO2-EOR also possible returns) to the CO2 transport sector.
co2 _ s EFF _ CO 2 OILPRICE h,,, n s a a TDh co2 _ s h,,,,,, n s a mu _ co 2 h n s a h co2 _ s VC _ CO 2 INTC _ S· co2 _ s2 hnsa,,,,,,,, nsa t hnsa STOR DFa PDa (7.18) a FC_ CO 2 inv_ co 2 _ s n,,,, s a n s aa aa USE_2C O s,, a aa INVC_ CO 2 inv _ co 2 _ s n,,,, s a n s aa aa USE_2 CO s,, a aa
Storage entities maximize their objective functions subject to a respective diffusion constraint which limits their maximal annual investment based on previous investments.
diff_ co 2 _ s 0START _ CO 2s inv _ co 2 _ s nsaa,,,,, DIFF_2 CO s inv _ co 2 _ s nsa sa 0 (7.19) n aa a n
Further constraints restrict the annual storage quantities based on prior investments as well as the overall maximal storage quantity per site and technology.
cap_ co 2 _ s 0 inv _ co 2 _ sn,s, aa co 2 _ s h , n ,s, a h , n ,s, a 0 (7.20) aaUSE_ CO2s,, a aa
151 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
max_ stor 0MAX _ STORns,,,,,, TD h PD aa co 2 _ s hnsaa nsa 0 (7.21) h aa a
7.2.7 Market clearing conditions across all sectors Three market clearing conditions connect the different sites (represented as nodes) and sectors in the
ELCO model: The first two represent the energy balance, while the third balances CO2 flows. With the introduction of the CfD scheme, the electricity market is fragmented: Technologies not supported by the CfD scheme market their generation to serve residual demand that remains after subtracting supply from CfD supported technologies shown in Equation (7.22). The free dual variable mu_eh,n,a of this equation corresponds to the price observed at the electricity wholesale market. By contrast, CfD technologies do not observe any feedback between their generation and market demand, just like in reality. Therefore, an additional curtailment constraint needs to be introduced in Equation (7.23), that limits total generation to meet the total demand.
0ghnta,,,,,,,,,,,,,,,,, g _ cfd hntaaa el _ t hnnna el _ t hnnna D hna RES _ OLD hna (7.22) t aa USE_ ELt,, a aa nn nn
mu _ ehna,, ( free ) h , n , a
0D RES _ OLD g g _ cfd curt_ g 0 (7.23) hna,,,,,,,,,,,, hna hnta hntaaa na n n t aa USE_ ELt,, a aa
The third market clearing is the CO2 flow balance with its free dual variable mu_CO2h,n,a.
0co 2 _ t co 2 _ s co 2 _ c g _ cfd EF CR _ G hnnna,,,,,,,,,,,,, hnsa hnia hntaaa t t nn s i t aa USE_ ELt,, a aa
co2 _ th,,,,, nn n a mu _ co 2 h n a ( free ) h , n , a (7.24) nn
7.3 Case study: the UK Electricity Market Reform The UK energy and climate policy used to be subject to a significant dichotomy between its policy targets and reality. Despite of fixed goals on final energy consumption from renewables (15% in 2020) and binding five-year carbon reduction targets towards a 80% reduction by 2050, the current energy policy framework was lacking instruments to incentivize investments that are necessary to achieve these goals. In addition, up to 20 GW of mostly coal fired generation have exceeded 40 years of age in the year 2015 and are either to be decommissioned or in need of retrofit investments. The upcoming decade therefore becomes vital for a future decarbonized electricity market to prevent stranded investments in carbon intensive power plants. The UK government decided to undertake a major restructuring of its energy policy framework, called Electricity Market Reform (EMR) (The Parliament of Great Britain 2013). The EMR introduces four main policies to support low-carbon technologies: Contracts for Differences (CfD), Carbon Floor Price (CFP), Emissions Performance Standards (EPS) and a Capacity Market (CM).
152 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
These instruments constitute a major reform to the previous framework of the UK electricity market which was characterized by a high competitiveness and low market concentration (DECC 2014b). Thus, its effects have been controversially discussed, e.g. by (Pollitt and Haney 2013; Chawla and Pollitt 2013). Some critics question the effect the reform might have on the UK electricity market and in particular on the future of low-carbon technologies. The future generation mix will be mostly determined by the government through long-term contracts with little ability to react quickly to future changes. Major risks include possible welfare losses as well as possible breached climate targets due to stranded investments in carbon intensive power plants (a topic examined by Johnson et al. (2015) on a global level). This calls for additional research on low-carbon technologies in the UK. Chalmers et al. (2013) summarize the findings of the two-year UKERC research project on the implementation of CCTS in the UK. To our best knowledge, however, there is no model that evaluates the effects of the UK-EMR on the UK electricity market as well as on the overall CCTS value chain including also the main industrial CO2 emitters.
The following section describes the UK-EMR and the policy measures which are included in the ELCO model.108 The used data set and results of this case study are afterwards discussed in the sections 7.3.2 and 7.3.3.
7.3.1 Describing the instruments: Contracts for Differences, Carbon Price Floor, and Emissions Performance Standard Contracts for Differences (CfD) were tied in the UK Energy Bill in 2013. They consist of a strike price for different low-carbon technologies resembling a fixed feed-in tariff. Generators take part in the normal electricity market but receive top-up payments from the government if the achieved prices are lower than the strike price. The government, on the other hand, receives equivalent payments from the generator if the market price exceeds the strike price. CfD and inherent strike prices are fixed for the duration of the contract. The long-term target of the CfD scheme is to find the most competitive carbon neutral technologies. In the short run, strike price levels are decided on in a technology-specific administrative negotiation process. In the long run, it is envisioned to determine a common strike price via a technology-neutral auction.
The UK government hopes that CfD enhance future investments as feed-in tariffs reduce the risk of market prices and gives incentives for cost reductions. Technologies that should be supported through CfD are various kinds of renewables (e.g. on-/offshore wind, PV, tidal, etc.) but also CCTS and nuclear. International dissent exists especially for the latter. Critics argue that a CfD for nuclear energy resembles an illegal subsidy tailored for the newly planned “Hinkley Point” project. The European Commission (EC) regulation requires implementation for an entire technology and accessibility for all possible investors. The nuclear sector, on the other hand, is due to its technology and safety specifics only open for a limited number of actors. The EC, however, decided in favour of the project after a
108 The specifics of a possible capacity market in the UK are not clear yet and were therefore not included in this case study.
153 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
formal investigation in October 2014, which might also have an effect on nuclear policies in other countries (Černoch and Zapletalová 2015).
The UK introduced a Carbon Price Floor (CPF) of 16 £/tCO2 (around 20 €/tCO2) for the electricity generators in 2013 to reduce uncertainty for investors. The CPF consists of the EU-ETS CO2 price and a variable climate change levy on top (carbon price support (CPS)). Forecasting errors in predicting the price of EU-ETS two years ahead can lead to distortions between the targeted and the final CPF. The climate change levy actually already exists since 2001, but the electricity sector used to be exempted from it. In 2013, the levy is expected to generate around £1 bn in the year 2013 (Ares 2014).
Initially, the CPF was planned to be gradually increasing to reach a target price of 30 £/tCO2 (around
38 €/tCO2) in 2020 and 70 £/tCO2 (around 88 €/tCO2) in 2030. A constantly rising minimum price should ensure increasing runtimes for low-carbon technologies such as renewables, nuclear and
CCTS as fossil based electricity generation becomes more expansive due to their CO2 emissions. The British minister for finance, however, announced in March 2014 that the CPF will be frozen at a level of
18 £/tCO2 (around 23 €/tCO2) until 2019/20 (G. Osborne 2014). The reason for this decision was the increasing discrepancy between the CPF and the EU-ETS CO2 emission price, lowering the competitiveness of British firms. It is yet unclear, how the CPF will evolve after 2020; depending probably largely on the effect of the upcoming structural reform of the EU-ETS. The CPS only has an effect on the British electricity sector. Neither is the combustion of natural gas for heating or cooking nor are electricity imports from neighboring countries affected by this instrument. The latter is also the main reason why the CPS has not been implemented in Northern Ireland which is part of the single electricity market in Ireland. (Pollitt and Haney 2013)
Another instrument implemented in the Energy Bill is the CO2 Emissions Performance Standard (EPS)
(The Parliament of Great Britain 2013). It limits the maximal annual CO2 emission of newly built or retrofitted electricity units to the ones of an average gas-fired power plant without carbon capture. Plants with higher carbon intensities like coal-fired units either have to reduce their load factor or install capture facilities for parts of their emissions. The EPS for a unit can be calculated by multiplying its capacity with 450 gCO2/kWh times 7,446 h (equivalent to a 0.85 load factor and 8,760h per year). This results in an annual CO2 budget of 3,350 tCO2/MW, restricting a coal-fired unit with emissions of 750 g/kWh to a maximal load factor of 0.5 or 4,470 h per year. The goal of this regulation is to foster investment in new gas power plants as well as power plants with capturing units. Power plants with capture units are additionally exempted from EPS for the first three years of operation to optimize their production cycles. Special exemptions exist for biomass emissions of plants below 50 MW related to heat production and in the case of temporary energy shortage.
154 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
7.3.2 Data input Electricity generation capacities as well as data for investment cost, variable cost, fixed cost, availability and life time assumptions are taken from DECC (2013a; 2014a). We assume a linear cost reduction over time for the investment cost according to Schröder et al. (2013); variable and fixed cost remain constant. The costs are independent from power plant location; but availabilities of renewables do vary. Industrial CO2 emissions and their location are taken from studies concentrating on CCTS adoption in the UK industry sector (Element Energy et al. 2014; Houses of Parliament
2012). Capturing costs in the industry sector as well as costs for CO2 storage and CO2-EOR application are taken from Mendelevitch
(2014). The fix costs are included in the variable capturing costs. Figure 7.1: Simplified network The simplified representation used for this case study consists of three nodes (see Figure 7.1). Node 1 and 2 represent the Northern and Southern part of the UK with their power plants and industrial facilities. A third offshore node resembles possible locations for offshore wind parks as well as CO2 storage with and without CO2-EOR in the North Sea. We assume electricity and CO2 pipeline connections between node 1 and 2 as well as between node 2 and node 3. We assume a simplified electricity grid neglecting congestion between nodes in this scenario. In addition, no exchange with the neighboring countries is allowed. CO2 pipelines can endogenously be constructed between adjacent nodes.
The CPF is assumed to remain constant at 18 £/tCO2 (around 23 €/tCO2) until 2020. We assume the
CO2 price to increase due to the effects of the structural reform of the EU-ETS. CPF and CO2 price are thus assumed to have the same level from 2030 onwards, rising linearly from €35 in 2030 to €80 in 2050. We include the given price projections for the strike prices in 2015 and 2020 DECC (2013b). These technology specific differences will be linearly reduced until 2030. Starting from 2030 all technologies under the CfD will be given the same financial support via an endogenous auctioning system. The EPS is set at a level of 450 g/kWh. An annual CO2 emissions reduction of 1% in the electricity sector is implemented leading to 90% emissions reduction in 2050 compared to 1990. No specific RES target is set. The discount rate is 5% for all players. The oil price is expected to remain at its current level of 65 €/bbl.
The annual load duration curve of UK is approximated by five weighted type hours, assuming a demand reduction of 20% till 2050 (base year 2015). This simplification does not allow for demand shifting nor energy storage in between type hours. CO2 emissions from industrial sources are assumed to decline by 40% until 2050. The lifetime of the existing power plant fleet varies by technology between 25 (most renewables), 40 (gas) and 50 (coal, nuclear, and hydro) years.
155 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
7.3.3 Case study results This simplified base case was created to show the characteristics and features of the ELCO model. Its results should not be over-interpreted but give an idea of the potential of the model, once its complete data set is calibrated.
The implementation of the various policy measures leads to a diversified electricity portfolio in 2050: with no specific RES target in place, renewables account for 46% of generation, gas (26%), nuclear (15%), and CCTS (13%). The majority of the investments in new renewable capacity happen before 2030. Less favorable regional potentials and technologies such as PV are only used in later periods. The implemented incentive mechanism is comparable to an auctioning system of “uniform pricing” where the last bidder sets the price. The average payments for low-carbon technologies are in the range of 80 to 110 €/MWh but depend strongly on the assumptions for learning curves and technology potentials. Different allocation mechanisms such as “pay as bid” might lower the overall system costs.
The share of coal-fired energy production is sharply reduced from 39% in 2015 to 0% in 2030 due to a phasing-out of the existing capacities (see Figure 7.2). New investments in fossil capacities occur for gas-fired CCGT plants, which are built from 2030 onwards. EPS hinders the construction of any new coal-fired power plant without CO2 capture. Sensitivity analysis shows that a change of its current level of 450 g/kWh in the range of 400-500 g/kWh has only little effect: Gas-fired power plants would still be allowed sufficient run-time hours while coal-fired plants remain strongly constrained. The overall capacity of nuclear power plants is slightly reduced over time.109 The share of renewables in the system grows continuously from 20% in 2015 to 30% in 2030 and 46% in 2050. Wind off- (41% in 2050) and onshore (25% in 2050) are the main renewable energy sources followed by hydro and biomass (together 27% in 2050).
CO2-EOR creates additional returns for CCTS deployment through oil sales. These profits trigger investments in CCTS regardless of additional incentives from the energy market. The potential for
CO2-EOR is limited and will be used to its full extent until 2050. The maximum share of CCTS in the electricity mix is 16% in 2045. The combination of assumed ETS and oil price also triggers CCTS deployment in the industry sector from 2020 onwards (see Figure 7.3). The industrial CO2 capture rate, contrary to the electricity sector, is constant over all type hours. The storage process requires a constant injection pressure, especially when connected to a CO2-EOR operation. This shows the need for intermediate CO2 storage to enable a continuous storage procedure and should be more closely examined in further studies. From 2030 onwards, emissions in the industrial sector are captured with the maximum possible capture rate of 90%. The usage of saline aquifers as well as depleted oil and gas fields is not beneficial assuming a CO2 certificate price of 80 €/tCO2 in 2050.
109 This is influenced through the diffusion constraint which limits the maximal annual construction, esp. in early periods.
156 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
Figure 7.2: Electricity generation (top) and power plant investment (bottom) from 2015-2050. Source: ELCO model results.
Figure 7.3: CO2 capture by electricity and industrial sector (area) and CO2 storage (bars) in 2015, 2030 and 2050 Source: Own modeling results with the ELCO model.
7.4 Conclusion: findings of an integrated electricity-CO2 modeling approach
This chapter presents a general electricity-CO2 modeling framework (ELCO model) that is able to simulate interactions of the energy-only market with different forms for national policy measures as well as a full representation of the carbon capture, transport, and storage (CCTS) chain. Different measures included in the model are feed-in tariffs, a minimum CO2 price and Emissions Performance Standards (EPS). Additionally, the model includes large point industrial emitters from the iron and steel
157 Chapter 7: The Impact of Policy Measures on Future Power Generation Portfolio and Infrastructure – A Combined Electricity and CCTS Investment and Dispatch Model (ELCO)
as well as cement sector that might also invest in carbon capture facilities, increasing scarcity for CO2 storage. Therefore, the modeling framework mimics the typical issues encountered in coal-based electricity systems that are now entering into transition to a low-carbon generation base. The model can be used to examine the effects of different envisioned policy measures and evaluate policy trade- off.
This chapter is used to describe the different features and potentials of the ELCO model. Such characteristics can easily be examined with a simplified model, even though its quantitative results should not be over-interpreted. As further development steps we need to test the robustness of the equilibrium results with sensitivity analysis while increasing the regional and time resolution of the model.
The results of the case study on the UK electricity market reform (EMR) present a show case of the model framework. It incorporates the unique combination of a fully represented CCTS infrastructure and a detailed representation of the electricity sector in UK. The instruments of the UK EMR, like EPS, CfD and CPF are integrated into the framework. Also we take into account demand variation in type hours, the availability of more and less favorable locations for RES and limits for their annual diffusion.
The model is driven by a CO2 target and an optional RES target.
The next steps are to compare the costs of different incentive schemes and to analyze their effects on the deployment of different low-carbon technologies, with a special focus on CCTS with and without the option for CO2-enhanced oil recovery (CO2-EOR). The role of industry CCTS needs to be further considered in this context. Additionally, we plan to study the feedback effects between the CfD scheme and the electricity price, and investigate the incentives of the government which acts along the three pillars of energy policy: cost-efficiency, sustainability and security; in a two-level setting. This also includes calculating the system integration costs of low-carbon technologies. A more detailed representation of the electricity transmission system operator (TSO) as market organizer helps doing so by separating financial and physical flows. The TSO is on the one hand responsible to guarantee supply meeting demand at any time and on the other hand reimburses CfD technologies for curtailment. At a later stage, we want to use the model for more realistic case studies to draw conclusions and possible policy recommendations for low-carbon support schemes in the UK as well as in other countries.
158 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
BIBLIOGRAPHY Abanades, J. C., B. Arias, A. Lyngfelt, T. Mattisson, D. E. Wiley, H. Li, M. T. Ho, E. Mangano, and S. Brandani. 2015. “Emerging CO2 Capture Systems.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 126–66. doi:10.1016/j.ijggc.2015.04.018. ADB. 2015. “Fossil Fuel Subsidies in Indonesia; Trends, Impacts, and Reforms.” Mandaluyong City, Philippines: Asian Development Bank. http://lib.myilibrary.com?id=902326. Aichele, Rahel, and Gabriel Felbermayr. 2012. “Kyoto and the Carbon Footprint of Nations.” Journal of Environmental Economics and Management 63 (3): 336–54. doi:10.1016/j.jeem.2011.10.005. ———. 2015. “Kyoto and Carbon Leakage: An Empirical Analysis of the Carbon Content of Bilateral Trade.” Review of Economics and Statistics 97 (1): 104–15. doi:10.1162/REST_a_00438. Ainger, D., S. Argent, and S. Haszeldine. 2010. “Feasibility Study for Europe-Wide CO₂ Infrastructures.” Study for the European Commission Directorate-General Energy Issue Rev.00 TREN/372-1/C3/2009. Leeds, UK: Ove Arup & Partners Limited. https://ec.europa.eu/energy/sites/ener/files/documents/2010_10_co2_infrastructures.pdf. Allevi, E., F. Bonenti, and G. Oggioni. 2013. “Compelementarity Models for Restructured Electricity Markets under Environmental Regulation.” Statistica & Applicazioni; Special Issue 2013. Italy. Anderson, Kym, and Warwick J. McKibbin. 2000. “Reducing Coal Subsidies and Trade Barriers: Their Contribution to Greenhouse Gas Abatement.” Environment and Development Economics, no. 04 (Oktober): 457–481. APEC. 2010. “2010 Leaders’ Declaration.” Asia-Pacific Economic Cooperation. November 13. http://www.apec.org/Meeting-Papers/Leaders-Declarations/2010/2010_aelm.aspx. Arabella Advisors. 2015. “Measuring the Growth of the Global Fossil Fuel Divestment and Clean Energy Investment Movement.” Arabella Advisors. Ares, Elena. 2014. “Carbon Price Floor.” London, UK: House of Commons Library..
ARI, and Mezler-Consulting. 2010. “Optimization of CO2 Storage in CO2 Enhanced Oil Recovery Projects.” London, UK: Department of Energy & Climate Change (DECC), Office of Cabon Capture and Storage. Arroyo-Currás, Tabaré, Nico Bauer, Elmar Kriegler, Valeria Jana Schwanitz, Gunnar Luderer, Tino Aboumahboub, Anastasis Giannousakis, and Jérôme Hilaire. 2015. “Carbon Leakage in a Fragmented Climate Regime: The Dynamic Response of Global Energy Markets.” Technological Forecasting and Social Change 90, Part A (January): 192–203. doi:10.1016/j.techfore.2013.10.002. Arze del Granado, Francisco Javier, David Coady, and Robert Gillingham. 2012. “The Unequal Benefits of Fuel Subsidies: A Review of Evidence for Developing Countries.” World Development 40 (11): 2234–48. doi:10.1016/j.worlddev.2012.05.005. Asheim, Geir B. 2013. “A Distributional Argument for Supply-Side Climate Policies.” Environmental and Resource Economics 56 (2): 239–54. doi:10.1007/s10640-012-9590-2. Ashworth, Peta, Sarah Wade, David Reiner, and Xi Liang. 2015. “Developments in Public Communications on CCS.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 449–58. doi:10.1016/j.ijggc.2015.06.002. Aune, Finn Roar, Rolf Golombek, Sverre A. C. Kittelsen, and Knut Einar Rosendahl. 2004. “Liberalizing the Energy Markets of Western Europe – a Computable Equilibrium Model Approach.” Applied Economics 36 (19): 2137–49. doi:10.1080/00036840310001641742.
159 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Australian Government. 2016. “Resources: Coal.” Australian Government Department of Industry, Innovation And Science. http://www.industry.gov.au/resource/Mining/ AustralianMineralCommodities/Pages/Coal.aspx. Babiker, Mustafa H. 2005. “Climate Change Policy, Market Structure, and Carbon Leakage.” Journal of International Economics 65 (2): 421–45. doi:doi:10.1016/j.jinteco.2004.01.003.
Bachu, Stefan. 2015. “Review of CO2 Storage Efficiency in Deep Saline Aquifers.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 188–202. doi:10.1016/j.ijggc.2015.01.007. Bárány, Ambrus, and Dalia Grigonytė. 2015. “Measuring Fossil Fuel Subsidies.” Brussels, Belgium: European Commission, Directorate-General for Economic and Financial Affairs. http://bookshop.europa.eu/uri?target=EUB:NOTICE:KCAY15040:EN:HTML. Baruya, Paul. 2007. “Supply Costs for Internationally Traded Coal.” London, UK: IEA Clean Coal Centre. Bast, Elizabeth, Alex Doukas, Sam Pickard, Laurie van der Burg, and Shelagh Whitley. 2015. “Empty Promises: G20 Subsidies to Oil, Gas and Coal Production.” Oversea Development Institute (ODI) and Oil Change International (OCI). http://www.qualenergia.it/sites/default/files/articolo- doc/Fossil_subsidies_full_ODI.pdf. Bauer, Nico, Ioanna Mouratiadou, Gunnar Luderer, Lavinia Baumstark, Robert J. Brecha, Ottmar Edenhofer, and Elmar Kriegler. 2013. “Global Fossil Energy Markets and Climate Change Mitigation – an Analysis with REMIND.” Climatic Change, October, 1–14. doi:10.1007/s10584- 013-0901-6. Bayer, Arne. 2012. “Investment Dynamics and Future Production Costs of Coal.” presented at the Session on: “Long-Term Costs and Reserves of Coal, Oil, and Natural Gas,” Berlin, Germany, March 22. Beaton, Christopher, Ivetta Gerasimchuk, Tara Laan, Kerryn Lang, Damon Vis-Dunbar, and Peter Wooders. 2013. “A Guidebook to Fossil-Fuel Subsidy Reform for Policy Makers in Southeast Asia.” Winnipeg, Canada: International Institute for Sustainable Development.
Bentham, M. S., K. L. Kirk, and J. Wiliams. 2008. “Basin-by-Basin Analysis of CO2 Storage Potential of All-Island Ireland.” Sustainable Energy and Geophysical Surveys Programm Commissioned Report CR/08/036. UK: British Geological Survey. Bentham, Michele. 2006. “An Assessment of Carbon Sequestration Potential in the UK – Southern North Sea Case Study.” Working Paper 85. Nottingham and Keyworth, UK: Tyndall Centre for Climate Change Research and British Geological Survey. BGR. 2015. “Energy Study 2015: Reserves, Resources and Availability of Energy Resources.” 19. Hannover, Germany: The Federal Institute for Geosciences and Natural Resources (BGR) on behalf of the German Mineral Resources Agency (DERA). http://www.bgr.bund.de/EN/Themen/Energie/Downloads/energiestudie_2013_en.pdf;jsessioni d=B54D487C8482BDF37835843002A0A66D.1_cid331?__blob=publicationFile&v=2.
Birkholzer, Jens T., Curtis M. Oldenburg, and Quanlin Zhou. 2015. “CO2 Migration and Pressure Evolution in Deep Saline Aquifers.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 203– 20. doi:10.1016/j.ijggc.2015.03.022. BMWi, and BMU. 2010. “Energy Concept - for an Environmentally Sound, Reliable and Affordable Energy Supply.” Berlin, Germany. http://www.bmwi.de/English/Redaktion/Pdf/energy- concept,property=pdf,bereich=bmwi,sprache=en,rwb=true.pdf. Böhringer, Christoph, Knut Einar Rosendahl, and Jan Schneider. 2014. “Unilateral Climate Policy: Can OPEC Resolve the Leakage Probem?” The Energy Journal 35 (4): 79–100. Boren, Zachary Davies. 2016. “China Stops Building New Coal-Fired Power Plants.” Energydesk. March 24. http://energydesk.greenpeace.org/2016/03/24/china-crackdown-new-coal-power- plants/.
160 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
BP. 2016. “Statistical Review of World Energy 2016.” 65 edition. London, UK: British Petroleum. http://www.bp.com/content/dam/bp/excel/energy-economics/statistical-review-2016/bp- statistical-review-of-world-energy-2016-workbook.xlsx. Britt, A.F., D. Summerfield, A. Whitaker, P. Kay, D.C. Champion, D. Huston, A.B. Senior, et al. 2015. “Australia’s Identified Mineral Resources 2015.” Australia’s Identified Mineral Resources. Geoscience Australia. http://www.ga.gov.au/metadata-gateway/metadata/record/87839/. Brook, M., K. Shaw, C. Vincent, and S. Holloway. 2009. “Gestco Case Study 2a-1: Storage Potential of the Bunter Sandstone in the UK Sector of the Southern North Sea and the Adjacent Onshore Area of Eastern England.” Keyworth, Nottingham, UK: Sustainable Energy and Geophysical Surveys Programm. Commissioned Report CR/03/154. Burniaux, Jean-Marc, and Jean Chateau. 2014. “Greenhouse Gases Mitigation Potential and Economic Efficiency of Phasing-out Fossil Fuel Subsidies.” International Economics 140 (December): 71–88. doi:10.1016/j.inteco.2014.05.002. Burniaux, Jean-Marc, and Joaquim Oliveira Martins. 2012. “Carbon Leakages: A General Equilibrium View.” Economic Theory 49 (2): 473–95. doi:10.1007/s00199-010-0598-y. Capros, Pantelis, L. Mantzos, D. Kolokotsas, N. Ioannou, T. Georgakopoulos, A. Filippopoulitis, and Y. Antoniou. 1998. “The PRIMES Energy System Model–reference Manual.” National Technical University of Athens, Greece. Černoch, Filip, and Veronika Zapletalová. 2015. “Hinkley Point C: A New Chance for Nuclear Power Plant Construction in Central Europe?” Energy Policy 83 (August): 165–68. doi:10.1016/j.enpol.2015.04.002. Chalmers, Hannah, Jon Gibbins, Rob Gross, Stuart Haszeldine, Phil Heptonstall, Florian Kern, Nils Markusson, Peter Pearson, Jim Watson, and Mark Winskel. 2013. “Analysing Uncertainties for CCS: From Historical Analogues to Future Deployment Pathways in the UK.” Energy Procedia, GHGT-11, 37: 7668–79. doi:10.1016/j.egypro.2013.06.712. Chandler, Alfred D. 1972. “Anthracite Coal and the Beginnings of the Industrial Revolution in the United States.” Business History Review 46 (02): 141–81. doi:10.2307/3113503. Chawla, Mallika, and Michael Pollitt. 2013. “Global Trends in Electricity Transmission System Operation: Where Does the Future Lie?” The Electricity Journal 26 (5): 65–71. doi:doi:10.1016/j.tej.2013.05.004. Chen, Yihsu, and Lizhi Wang. 2013. “Renewable Portfolio Standards in the Presence of Green Consumers and Emissions Trading.” Networks and Spatial Economics 13 (2): 149–81. doi:10.1007/s11067-012-9176-0. Clements, Benedict J., David Coady, Stefania Fabrizio, Sanjeev Gupta, Trevor Alleyne, and Carlo Sdralevich, eds. 2013. Energy Subsidy Reform: Lessons and Implications. Washington, D.C: International Monetary Fund. cmpdi. 2014. “UNFC (2004) Classification of Coal Resources of CIL Blocks as on 01.04.2013.” Ranchi, India: Central Mine Planning and Design Institute Limited. Coady, David, Ian Parry, Louis Sears, and Baoping Shang. 2015. “How Large Are Global Energy Subsidies?” IMF Working Paper WP/15/105. Washington, DC, USA: International Monetary Fund, Fiscal Affairs Department. Coal Controller’s Organisation. 2015. “Coal Directory of India 2013 – 2014 Coal Statistics.” Kolkata, India: Government of India, Ministry of Coal, Coal Controller’s Organisation. http://coal.nic.in/sites/upload_files/coal/files/coalupload/coaldir13-14.pdf. Cole, Daniel H. 2015. “Advantages of a Polycentric Approach to Climate Change Policy.” Nature Climate Change 5 (2): 114–18. doi:10.1038/nclimate2490. Collier, Paul, and Anthony J. Venables. 2014. “Closing Coal: Economic and Moral Incentives.” Oxford Review of Economic Policy 30 (3): 492–512. doi:10.1093/oxrep/gru024. Collins, Kim, and Roman Mendelevitch. 2015. “Leaving Coal Unburned: Options for Demand-Side and Supply-Side Policies.” 87. DIW Roundup. Berlin, Germany: German Institute for Economic Research (DIW Berlin). http://www.diw.de/documents/publikationen/73/diw_01.c.526889.de/diw_roundup_91_en.pdf. 161 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Commonwealth of Australia. 2015. “Coal in India.” http://www.industry.gov.au/Office-of-the-Chief- Economist/Publications/Pages/Coal-in-India.aspx. Cournot, Antoine Augustin. 1838. Recherches Sur Les Principes Mathématiques de La Théorie Des Richesses. 1. Chez L. Hachette, Libraire de l´Université Royale de France, Paris. Cramton, Peter, Axel Ockenfels, and Steven Stoft. 2015. “An International Carbon-Price Commitment Promotes Cooperation.” Economics of Energy & Environmental Policy 4 (2). doi:10.5547/2160-5890.4.2.aock. Criqui, Patrick. 1996. “International Markets and Energy Prices: The POLES Model.” In Models for Energy Policy, edited by Jean-Baptiste Lesourd, Jacques Percebois, and François Valette. Routledge Studies in the History of Economic Modelling 1. London ; New York: Routledge. Crocker, Geoff, and Alex Kovalchuk. 2008. Prospects for Coal and Clean Coal Technologies in Russia. IEA Coal Research, Clean Coal Centre. International Energy Agency Coal Research. https://books.google.de/books?id=VT5HPgAACAAJ. Dahl, Carol. 1993. “A Survey of Energy Demand Elasticities in Support of the Development of the NEMS.” MPRA Paper. University Library of Munich, Germany. http://econpapers.repec.org/paper/pramprapa/13962.htm. Dansie, Grant, Marc Lanteigne, and Indra Overland. 2010. “Reducing Energy Subsidies in China, India and Russia: Dilemmas for Decision Makers.” Sustainability 2 (2): 475–93. doi:10.3390/su2020475. Dartanto, Teguh. 2013. “Reducing Fuel Subsidies and the Implication on Fiscal Balance and Poverty in Indonesia: A Simulation Analysis.” Energy Policy 58 (July): 117–34. doi:10.1016/j.enpol.2013.02.040. Davidson, C., R. Dahowski, and J. Dooley. 2011. “A Quantitative Comparison of the Cost of Employing EOR-Coupled CCS Supplemented with Secondary DSF Storage for Two Large CO2 Point Sources.” Energy Procedia 4 (January): 2361–68. DECC. 2013a. “Electricity Generation Costs.” London, UK: Department of Energy & Climate Change (DECC). ———. 2013b. “Investing in Renewable Technologies - CfD Contract Terms and Strike Prices.” London, UK: Department of Energy & Climate Change (DECC). ———. 2014a. “Digest of United Kingdom Energy Statistics 2014.” A Nationals Statistics Publication. London, UK: Department of Energy & Climate Change (DECC). ———. 2014b. “UK Energy in Brief 2014.” A National Statistics Publication. London, UK: Department of Energy & Climate Change (DECC). Denjean, Benjamin, Ivetta Gerasimchuk, Ken Bossong, Sam Pickard, and Shelagh Whitley. 2015. “G20 Subsidies to Oil, Gas and Coal Production: China.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/9985.pdf. Denniss, Richard. 2015a. “When You’re in a Hole - Stop Digging! The Economic Case for a Moratorium on New Coal Mines.” Discussion Paper. The Australian Institute. http://www.tai.org.au/sites/defualt/files/Denniss%202015%20- %20When%20you%20are%20in%20a%20hole%20stop%20digging_0.pdf. ———. 2015b. “Moratorium on Coal Mines Makes Sense for All.” Australian Financial Review. November 3. http://www.afr.com/opinion/moratorium-on-coal-mines-makes-sense-for-all- 20151102-gkojop. DERA. 2012. “Energy Study 2012: Reserves, Resources and Availability of Energy Resources.” Hannover, Germany: The Federal Institute for Geosciences and Natural Resources (BGR) on behalf of the German Mineral Resources Agency (DERA). Di Maria, Corrado, and Edwin van der Werf. 2008. “Carbon Leakage Revisited: Unilateral Climate Policy with Directed Technical Change.” Environmental and Resource Economics 39 (2): 55– 74. doi:10.1007/s10640-007-9091-x.
162 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
DOE/EIA. 2012. “Annual Energy Outlook 2012 - Energy Information Administration.” Washington DC, USA: Department of Energy, U.S. Energy Information Administration. http://www.eia.gov/forecasts/aeo/. Dooley, J. J., R. T. Dahowski, C. L. Davidson, M. A. Wise, N. Gupta, and S. H. Kim. 2006. “Carbon Dioxide Capture and Geological Storage - A Core Element of a Global Energy Technology Strategy to Address Climate Change.” Technology Report. College Park, USA: Global Energy Technology Strategy Program (GTSP). Doukas, Alex, and Shelagh Whitley. 2015. “G20 Subsidies to Oil, Gas and Coal Production: United States.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/publications/10086-g20-subsidies-oil-gas-coal-production-united- states. Downie, Christian. 2015. “Prolonged International Environmental Negotiations: The Roles and Strategies of Non-State Actors in the EU.” International Environmental Agreements: Politics, Law and Economics, August, 1–17. doi:10.1007/s10784-015-9292-7. DTEK. 2014. “DTEK 2014 Integrated Report - Financial and Non-Financial Performance.” Durand-Lasserve, Olivier, Lorenza Campagnolo, Jean Chateau, and Rob Dellink. 2015. “Modelling of Distributional Impacts of Energy Subsidy Reforms: An Illustration with Indonesia.” OECD Environment Working Papers 86. Paris, France: OECD Publishing. http://www.oecd- ilibrary.org/environment/modelling-of-distributional-impacts-of-energy-subsidy- reforms_5js4k0scrqq5-en. Eberhard, Anton. 2015. “Market, Investment, and Policy Challenges for South African Coal.” In The Global Coal Market: Supplying the Major Fuel for Emerging Economies, edited by Mark C. Thurber and Richard K. Morse. Cambridge, UK: Cambridge University Press. EC. 2009. European Energy Programme for Recovery. ———. 2010. “COMMISSION DECISION of 3 November 2010 Laying down Criteria and Measures for the Financing of Commercial Demonstration Projects That Aim at the Environmentally Safe Capture and Geological Storage of CO2 as Well as Demonstration Projects of Innovative Renewable Energy Technologies under the Scheme for Greenhouse Gas Emission Allowance Trading within the Community Established by Directive 2003/87/EC of the European Parliament and of the Council.” 2010/670/EU. Brussels, Belgium: European Commission. ———. 2011. “Energy Roadmap 2050.” Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions COM(2011) 885 final. Brussels, Belgium: European Commission. http://eur- lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2011:0885:FIN:EN:PDF. ———. 2012. “23 Innovative Renewable Energy Demonstration Projects Receive €1.2 Billion EU Funding.” http://ec.europa.eu/clima/news/articles/news_2012121801_en.htm. ———. 2013a. “Long Term Infrastructure Vision for Europe and beyond.” COM(2013) 711. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. Brussels, Belgium: European Commission. http://ec.europa.eu/energy/infrastructure/pci/doc/com_2013_0711_en.pdf. ———. 2013b. “EU Energy, Transport and GHG Emissions Trends to 2050: Reference Scenario 2013.” Brussels, Belgium: European Commission. http://ec.europa.eu/transport/media/publications/doc/trends-to-2050-update-2013.pdf. ———. 2014a. “A Policy Framework for Climate and Energy in the Period from 2020 up to 2030.” Commission Staff Working Document Impact Assessment COM(2014) 015 final. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. Brussels, Belgium: European Commission. http://eur-lex.europa.eu/legal- content/EN/TXT/?uri=CELEX%3A52014DC0015.
163 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
———. 2014b. “Questions and Answers on the Proposed Market Stability Reserve for the EU Emissions Trading System.” Brussels, Belgium: European Commission. http://europa.eu/rapid/press-release_MEMO-14-39_en.pdf. Ecofys. 2014. “Subsidies and Costs of EU Energy.” Final Report. Project number: DESNL14583. By Order of European Commission. London, UK: Ecofys. Edenhofer, O., R. Pichs-Madruga, Y. Sokona, S. Kadner, J. C. Minx, S. Brunner, S. Agrawala, et al. 2014. “Technical Summary.” In Climate Change 2014: Mitigation of Climate Change. IPCC Working Group III Contribution to AR5. Cambridge University Press. http://www.ipcc.ch/pdf/assessment-report/ar5/wg3/ipcc_wg3_ar5_technical-summary.pdf. Edmond, Jae, and John M. Reilly. 1985. Global Energy: Assessing the Future. New York, USA: Oxford University Press. EEA. 2011. “The European Pollutant Release and Transfer Register.” Copenhagen, Denmark: European Environmental Agency. EFA. 2008. “General Scheme for the Positioning of Power Plants until 2020.” Energy Forecasting Agency (EFA), Available in Russian only at http://e-apbe.ru/scheme/gs.doc. Egerer, Jonas, Roman Mendelevitch, and Christian von Hirschhausen. 2014. A Lower Carbon Strategy for the Electricity Sector of Kazakhstan to 2030/50: Scenarios for Generation and Network Development ; Technical Report. Berlin: DIW Berlin. http://nbn- resolving.de/urn:nbn:de:0084-diwkompakt_2014-0857. Egging, Ruud, Franziska Holz, and Steven A. Gabriel. 2010. “The World Gas Model: A Multi-Period Mixed Complementarity Model for the Global Natural Gas Market.” Energy 35 (10): 4016–29. doi:10.1016/j.energy.2010.03.053. Ehrenmann, Andreas, and Yves Smeers. 2011. “Generation Capacity Expansion in a Risky Environment: A Stochastic Equilibrium Analysis.” Operations Research 59 (6): 1332–46. doi:10.1287/opre.1110.0992. EIA. 2008. “Annual Coal Report 2007.” Washington DC, USA: Department of Energy, U.S. Energy Information Administration. Available at http://www.eia.doe.gov/cneaf/coal/page/acr/acr_sum.html. ———. 2010. “Annual Coal Distribution Report.” Washington DC, USA: Department of Energy, U.S. Energy Information Administration. ———. 2011. “Rail Coal Transportation Rates to the Electric Power Sector 2001–2008.” Washington DC, USA: Department of Energy, U.S. Energy Information Administration. ———. 2012. “Annual Coal Report.” Washington DC, USA: Department of Energy, U.S. Energy Information Administration. ———. 2015a. “Direct Federal Financial Interventions and Subsidies in Energy in Fiscal Year 2013.” Independent Statistics & Analysis. Washington DC, USA: US Energy Information Administration, US Department of Energy. ———. 2015b. “Analysis of the Impacts of the Clean Power Plan.” Washington, D.C., USA: U.S. Energy Information Administration. https://www.eia.gov/analysis/requests/powerplants/cleanplan/pdf/powerplant.pdf. ———. 2015c. “Coal Mine Starts Continue to Decline.” September 23. http://www.eia.gov/todayinenergy/detail.cfm?id=23052. ———. 2016a. “International Energy Outlook 2016.” U.S. Energy Information Administration Office of Energy Analysis U.S. Department of Energy. http://www.eia.gov/forecasts/ieo/pdf/0484(2016).pdf. ———. 2016b. “Annual Coal Report 2014.” Independent Statistics & Analysis. Washington, D.C., USA: Department of Energy, U.S. Energy Information Administration.
Eide, Jan, Fernando J. de Sisternes, Howard J. Herzog, and Mort D. Webster. 2014. “CO2 Emission Standards and Investment in Carbon Capture.” Energy Economics 45 (September): 53–65. doi:10.1016/j.eneco.2014.06.005.
164 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Eisenack, Klaus, Ottmar Edenhofer, and Matthias Kalkuhl. 2012. “Resource Rents: The Effects of Energy Taxes and Quantity Instruments for Climate Protection.” Energy Policy, Special Section: Frontiers of Sustainability, 48 (September): 159–66. doi:10.1016/j.enpol.2012.05.001.
Element Energy, PSCE, Imperial College, and University of Sheffield. 2014. “Demonstrating CO2 Capture in the UK Cement, Chemicals, Iron and Steel and Oil Refining Sectors by 2025: A Techno-Economic Study.” Final Report for DECC and BIS. Cambridge, UK. Ellerman, Denny A., Vanessa Valero, and Aleksandar Zaklan. 2015. “An Analysis of Allowance Banking in the EU ETS.” Working Paper. RSCAS Working Papers. Florence, Italy: Florence School of Regulation. http://cadmus.eui.eu//handle/1814/35517. EMF. 1978. “Coal in Transition: 1980–2000.” Stanford, USA: Stanford University: Energy Modelling Forum. Energy Watch Group. 2007. “Coal: Resources and Future Production.” EWG-Paper #1/07. Available from http://energywatchgroup.org/. EPA. 2012. National Emission Standards for Hazardous Air Pollutants From Coal- and Oil-Fired Electric Utility Steam Generating Units and Standards of Performance for Fossil-Fuel-Fired Electric Utility, Industrial-Commercial-Institutional, and Small Industrial- Commercial- Institutional Steam Generating Units. 40 CFR 60 and Part 63. RIN 2060–AS39. United States Environmental Protection Agency. USA. ———. 2015a. “Coal Mine Methane Country Profiles - Chapter 8: Colombia.” Washington, D.C., USA: Environmental Protection Agency. http://www.globalmethane.org/documents/Toolsres_coal_overview_ch8.pdf. ———. 2015b. “EPA Fact Sheet: Social Costs of Carbon.” Washington, D.C., USA: Environmental Protection Agency. https://www3.epa.gov/climatechange/Downloads/EPAactivities/social-cost- carbon.pdf. Erickson, Peter, and Michael Lazarus. 2016. “How Would Phasing out U.S. Federal Leases for Fossil Fuel Extraction Affect CO2 Emissions and 2°C Goals?” https://www.sei- international.org/publications?pid=2937. ExxonMobil. 2016. “The Outlook for Energy: A View to 2040.” Irving, USA: Exxon Mobil Corporation. http://cdn.exxonmobil.com/~/media/global/files/outlook-for-energy/2016/2016-outlook-for- energy.pdf. Fæhn, Taran, Cathrine Hagem, Lars Lindholt, Ståle Mæland, and Einar Rosendahl. 2017. “Climate Policies in a Fossil Fuel Producing Country - Demand versus Supply Side Policies.” The Energy Journal 38 (1). doi:10.5547/01956574.38.1.tfae. Fais, Birgit, Markus Blesl, Ulrich Fahl, and Alfred Voß. 2014. “Comparing Different Support Schemes for Renewable Electricity in the Scope of an Energy Systems Analysis.” Applied Energy 131: 479–89. doi:10.1016/j.apenergy.2014.06.046.
Felder, Stefan, and Thomas F. Rutherford. 1993. “Unilateral CO2 Reductions and Carbon Leakage: The Consequences of International Trade in Oil and Basic Materials.” Journal of Environmental Economics and Management 25 (2): 162–76. doi:10.1006/jeem.1993.1040. Fernandes, Ashish, and Tom Sanzillo. 2013. “Coal India Running on Empty? CIL Misleading Investors, Extractable Reserves Could Be Exhausted in 17 Years.” India: Greenpeace India. http://www.greenpeace.org/india/Global/india/report/2013/Coal-India-Running-on-Empty.pdf. Fernihough, Alan, and Kevin Hjortshøj O’Rourke. 2014. “Coal and the European Industrial Revolution.” Cambridge, MA: National Bureau of Economic Research. http://www.nber.org/papers/w19802.pdf. Finighan, Reuben. 2016. “The Case for a Coal Mine Moratorium: Reserves within Existing Mines versus the Carbon Budget.” Briefing Paper. Melbourne, Australia: University of Melbourne. Finkenrath, M. 2011. “Cost and Performance of Cabon Dioxide Capture from Power Generation.” International Energy Agency (IEA/OECD). Paris, France. Finon, D. 1979. “Scope and Limitations of Formalized Optimization of a National Energy System—the EFOM Model.” Energy Models for the European Community, an Energy Policy Special
165 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Published by IPC Science and Technology Press Limited for the Commission of the European Community, Brussels, Belgium.
Fischedick, Manfred, Klaus Görner, and Margit Thomeczek. 2015. CO2: Abtrennung, Speicherung, Nutzung. Berlin, Heidelberg, Germany: Springer. http://link.springer.com/10.1007/978-3-642- 19528-0. Fishbone, Leslie G., and Harold Abilock. 1981. “Markal, a Linear-Programming Model for Energy Systems Analysis: Technical Description of the Bnl Version.” International Journal of Energy Research 5 (4): 353–375. doi:10.1002/er.4440050406. Forsythe, Michael. 2016. “China Curbs Plans for More Coal-Fired Power Plants.” The New York Times, April 25. http://www.nytimes.com/2016/04/26/business/energy-environment/china- coal.html. Fortuny-Amat, J., and B. McCarl. 1981. “A Representation and Economic Interpretation of a Two-Level Programming Problem.” The Journal of the Operational Research Society 32 (9): 783–92.
Fritze, Kevin. 2009. “Modeling CO2 Storage Pipeline Routes in the United States.” Durham, USA: Nicholas School of the Environment and Earth Sciences. http://hdl.handle.net/10161/985. Fulton, M., T. Buckley, D. Koplow, L. Sussams, and A. Grant. 2015. “A Framework for Assessing Thermal Coal Production Subsidies.” London, UK: Energy Transition Advisors, IEEFA, Earth Track, Carbon Tracker Initiative. http://www.carbontracker.org/wp- content/uploads/2015/09/Thermal-Coal-Prod-Subsidies-final-12-9.pdf. Fulton, M., P. Spedding, R. Schuwerk, and L. Sussams. 2015. “The Fossil Fuel Transition Blue Print.” Energy Transition Advisors and Carbon Tracker Initiative. G20. 2009. “Leaders’ Statement the Pittsburgh Summit.” G20. http://www.g20.org/English/Documents/PastPresidency/201512/t20151225_1724.html. Gabriel, Steven A., and Florian U. Leuthold. 2010. “Solving Discretely-Constrained MPEC Problems with Applications in Electric Power Markets.” Energy Economics 32 (1): 3–14. doi:10.1016/j.eneco.2009.03.008. Gale, J., J. C. Abanades, S. Bachu, and C. Jenkins. 2015a. “Special Issue Commemorating the 10th Year Anniversary of the Publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 1–5. doi:10.1016/j.ijggc.2015.06.019. ———. 2015b. “Special Issue Commemorating the 10th Year Anniversary of the Publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 1–5. doi:10.1016/j.ijggc.2015.06.019. Gandhi, Raina, Christopher R. Knittel, Paula Pedro, and Catherine Wolfram. 2016. “Running Randomized Field Experiments for Energy Efficiency Programs: A Practitioner’s Guide.” Economics of Energy & Environmental Policy 5 (2). doi:10.5547/2160-5890.5.2.rgan. Garg, Vibhuti, and Ken Bossong. 2015. “G20 Subsidies to Oil, Gas and Coal Production: India.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/9959.pdf. Garg, Vibhuti, and Lucy Kitson. 2015. “G20 Subsidies to Oil, Gas and Coal Production: South Africa.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/9982.pdf. Gatot, Bambang, and Pandu Sjahrir. 2015. “Government of Indonesia Postpones the Coal Royalty Hike.” Indonesia Investments. July 23. http://www.indonesia-investments.com/news/todays- headlines/government-of-indonesia-postpones-the-coal-royalty-hike/item5776.
166 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
GCCSI. 2011. “Accelerating the Uptake of CCS: Industrial Use of Captured Carbon Dioxide.” Camberra, Australia: Global CCS Institute. ———. 2012. “CCS Ready Policy and Regulations - the State of Play.” Camberra, Australia: Global CCS Institute. ———. 2014. “The Global Status of CCS: 2014.” Canberra, Australia: Global CCS Institute. http://www.globalccsinstitute.com/publications/global-status-ccs-2014. ———. 2015a. “Global Storage Readiness Assessment: An Approach to Assessing National Readiness for Wide-Scale Deployment of CO2 Geological Storage Projects.” Camberra, Australia: Global CCS Institute. ———. 2015b. “The Global Status of CCS: 2015.” Camberra, Australia: Global CCS Institute. http://www.globalccsinstitute.com/publications/global-status-ccs-2015. GeoCapacity, E. U. 2009. “Assessing European Capacity for Geological Storage of Carbon Dioxide.” EU GeoCapacity Project no. SES6-518318. Final Report. Denmark: EU GeoCapacity - Geological Survey of Denmark and Greenland. Geosience Australia. 2014. “Black Coal.” Australian Government Geosience Australia. May 29. http://www.ga.gov.au/data-pubs/data-and-publications-search/publications/aimr/black-coal. Gerlagh, Reyer. 2011. “Too Much Oil.” CESifo Economic Studies 57 (1): 79–102.
Gerling, Johannes Peter. 2010. “CO2 Storage - German and International Perspective.” Presentation at Berlin Seminar on Energy and Climate Policy. DIW Berlin, Germany. Geske, Joachim, Niels Berghout, and Machteld van den Broek. 2015a. “Cost-Effective Balance between CO2 Vessel and Pipeline Transport. Part I – Impact of Optimally Sized Vessels and Fleets.” International Journal of Greenhouse Gas Control 36: 175–88. doi:10.1016/j.ijggc.2015.01.026.
———. 2015b. “Cost-Effective Balance between CO2 Vessel and Pipeline Transport: Part II – Design of Multimodal CO2 Transport: The Case of the West Mediterranean Region.” International Journal of Greenhouse Gas Control 33 (February): 122–34. doi:10.1016/j.ijggc.2014.12.005. Gilbert, Alexander Q., and Benjamin K. Sovacool. 2016. “Looking the Wrong Way: Bias, Renewable Electricity, and Energy Modelling in the United States.” Energy 94 (January): 533–41. doi:10.1016/j.energy.2015.10.135. Glencore. 2016. “Glencore Resources & Reserves as at 31. December 2015.” Glencore. http://www.glencore.com/assets/investors/doc/reports_and_results/2015/GLEN-2015- Resources-Reserves-Report.pdf. Gollier, Christian, and Jean Tirole. 2015. “Negotiating Effective Institutions against Climate Change.” Economics of Energy & Environmental Policy 4 (2). doi:10.5547/2160-5890.4.2.cgol.
Keefe, and Sarah Mander. 2014. “Public Perceptions of CO2 Transportation in׳Gough, Clair, Laura O Pipelines.” Energy Policy 70 (July): 106–14. doi:10.1016/j.enpol.2014.03.039.
Gozalpour, F., S. Ren, and B. Tohidi. 2005. “CO2 EOR and Storage in Oil Reservoir.” Oil & Gas Science and Technology 60 (3): 537–46. doi:10.2516/ogst:2005036. Grantham Research Institute. 2015a. “The Global Climate Legislation Database | Grantham Research Institute on Climate Change and the Environment.” The Global Climate Legislation Database. http://www.lse.ac.uk/GranthamInstitute/legislation/the-global-climate-legislation-database/. ———. 2015b. “Nicholas Stern Welcomes Initiative on Coal Mines by Anote Tong, President of the Republic of Kiribati | Grantham Research Institute on Climate Change and the Environment.” August 13. http://www.lse.ac.uk/GranthamInstitute/news/nicholas-stern-welcomes-initiative- on-coal-mines-by-anote-tong-president-of-the-republic-of-kiribati/. Gray, Matthew. 2015. “Coal: Caught in the EU Utility Death Spiral.” Carbon Tracker. http://www.carbontracker.org/wp-content/uploads/2015/06/CTI-EU-Utilities-Report-v6- 080615.pdf.
Greenpeace. 2011. “Potentielle CO2-Endlager in Deutschland in Salzwasser Führenden Tiefengestein.” Auswertung der BGR-Daten durch Greenpeace. Stand 01/2011, bearbeitet von: Karsten Smid. Hamburg, Germany: Greenpeace Germany e.V. 167 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Groenenberg, Heleen, and Heleen de Coninck. 2008. “Effective EU and Member State Policies for Stimulating CCS.” International Journal of Greenhouse Gas Control, TCCS-4: The 4th Trondheim Conference on CO2 Capture, Transport and Storage, 2 (4): 653–64. doi:10.1016/j.ijggc.2008.04.003. GSI. 2010. “Inventory of Geological Resource of Indian Coal.” Geological Survey of India. Available at http://www.cmpdi.co.in/docfiles/coalreserve_010410.pdf. ———. 2011. “The Friends of Fossil Fuel Subsidy Reform: Supporting the G-20 and APEC Commitments | Global Subsidies Initiative.” Global Subsidies Initiative. February 7. http://www.iisd.org/gsi/news/friends-fossil-fuel-subsidy-reform-supporting-g-20-and-apec- commitments. Gürkan, Gül, and Romeo Langestraat. 2014. “Modeling and Analysis of Renewable Energy Obligations and Technology Bandings in the UK Electricity Market.” Energy Policy 70 (July): 85–95. doi:10.1016/j.enpol.2014.03.022. Hackett, S. 1995. “Pollution-Controlling Innovation in Oligopolistic Industries: Some Comparisons between Patent Races and Research Joint Ventures.” Journal of Environmental Economics and Management 29 (3): 339–56. doi:10.1006/jeem.1995.1051. Haftendorn, Clemens. 2012. “Evidence of Market Power in the Atlantic Steam Coal Market Using Oligopoly Models with a Competitive Fringe.” Discussion Papers of DIW Berlin 1185. DIW Berlin, German Institute for Economic Research. Haftendorn, Clemens, Christian von Hirschhausen, and Franziska Holz. 2008. “Moving towards a ‘COAL-PEC’?” Weekly Report 4 (10): 62–67. Haftendorn, Clemens, and Franziska Holz. 2010. “Modeling and Analysis of the International Steam Coal Trade.” The Energy Journal 31 (4): 205–29. doi:10.5547/ISSN0195-6574-EJ-Vol31-No4- 10. Haftendorn, Clemens, Franziska Holz, and Christian von Hirschhausen. 2012. “The End of Cheap Coal? A Techno-Economic Analysis until 2030 Using the COALMOD-World Model.” Fuel 102 (December): 305–25. doi:10.1016/j.fuel.2012.04.044. Haftendorn, Clemens, Franziska Holz, Claudia Kemfert, and Christian von Hirschhausen. 2013. “Global Steam Coal Markets until 2030: Perspectives on Production, Trade and Consumption under Increasing Carbon Constraints.” In Handbook on Energy and Climate Change, S. 103- 122. Cheltenham [a.o.] : Elgar. Haftendorn, Clemens, Franziska Holz, Tim Winke, and Christian von Hirschhausen. 2011. “Entwicklungen Des Internationalen Handels Mit Kesselkohle: Lehren Aus Dem COALMOD- World-Modell.” Glückauf 147 (12): 566–71. ———. 2012. “Perspektiven Des Globalen Kesselkohlenmarktes: Das COALMOD-World-Modell.” 62 (2012), 3, S. 38-43 62 (3): S. 38-43. Haftendorn, Clemens, Claudia Kemfert, and Franziska Holz. 2012. “What about Coal? Interactions between Climate Policies and the Global Steam Coal Market until 2030.” Energy Policy 48: 274–83. doi:doi:10.1016/j.enpol.2012.05.032. Hagem, C., and H. Storrøsten. 2016. “Supply versus Demand-Side Policies in the Presence of Carbon Leakage and the Green Paradox.” 836. Discussion Papers. Oslo, Norway: Statistics Norway. Hagen, Achim, Leonhard Kähler, and Klaus Eisenack. 2016. “Transnational Environmental Agreements with Heterogeneous Actors.” Working Paper V-387-16. University of Oldenburg, Department of Economics. https://ideas.repec.org/p/old/dpaper/387.html. Hahn, Robert, and Robert Metcalfe. 2016. “The Impact of Behavioral Science Experiments on Energy Policy.” Economics of Energy & Environmental Policy 5 (2). doi:10.5547/2160-5890.5.2.rhah. Hamming, Richard W. 1962. Numerical Methods for Scientists and Engineers. New York, USA: McGraw-Hill. Harstad, Bård. 2012. “Buy Coal! A Case for Supply-Side Environmental Policy.” Journal of Political Economy 120 (1): 77–115. doi:10.1086/665405.
168 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Hartley, Peter, and Kenneth B. III Medlock. 2006. “The Baker Institute World Gas Trade Model.” In Natural Gas and Geopolitics, edited by David G. Victor, Amy M. Jaffe, and Mark H. Hayes, 357–406. Cambridge: Cambridge University Press. Hazeldine, Staurt. 2009. “Carbon Capture and Storage: UK’s Fourth Energy Pillar, or Broken Bridge?” Working Paper SCCS 2009-03. Edinburgh, UK: Scottish Centre for Carbon Storage.
Hecking, Harald. 2016. “Twenty-First Century, the Century of Coal? CO2 Prices to Curb Coal Demand.” Oxford Review of Economic Policy 32 (2): 260–81. doi:10.1093/oxrep/grw011.
Heddle, G., H. Herzog, and M. Klett. 2003. “The Economics of CO2 Storage.” MIT LFEE 2003-003 R. Cambridge, Mass, USA: Massachusetts Institute of Technology. Herold, Johannes, Christian von Hirschhausen, and Sophia Ruester. 2010. “Carbon Capture, Transport and Storage in Europe: A Problematic Energy Bridge to Nowhere?” CEPS Working Document 341. Brussels, Belgium: Centre for European Policy Studies (CEPS). Herold, Johannes, Pao-Yu Oei, and Andreas Tissen. 2011. “Ökonomische Aspekte von CCTS Im Industriesektor: Potentialabschätzung, Infrastruktur Und Nutzungskonkurrenzen.” WIP, TU Berlin, Germany: Gutachten im Auftrag des WWF Deutschland. Herold, Johannes, Sophia Rüster, and Christian von Hirschhausen. 2010. “Vertical Integration and Market Structure along the Extended Value Added Chain Including Carbon Capture, Transport and Sequestration (CCTS).” Project No 213744. Project Report for the European Commission SECURE Work Package 5.3.2. Dresden, Germany: Chair of Energy Economics and Public Sector Management, Dresden University of Technology. Hill, Taylor. 2016. “More EU Countries Are Giving Coal the Boot.” TakePart. April 7. http://www.takepart.com/article/2016/04/07/quarter-european-countries-ditch-coal. Hirschhausen, Christian von, Clemens Haftendorn, Johannes Herold, Franziska Holz, Anne Neumann, and Sophia Rüster. 2011. “European Supply Security with Coal - High Uncertainties due to Obstacles to Carbon Capture, Transport and Storage (CCTS).” CEPS Policy Brief. Brussels. Hirschhausen, Christian von, Johannes Herold, and Pao-Yu Oei. 2012a. “How a ‘Low Carbon’ Innovation Can Fail – Tales from a ‘Lost Decade’ for Carbon Capture, Transport, and Sequestration (CCTS).” Economics of Energy & Environmental Policy 1 (2): 115–23. doi:10.5547/2160-5890.1.2.8. ———. 2012b. “How a ‘Low Carbon’ Innovation Can Fail – Tales from a ‘Lost Decade’ for Carbon Capture, Transport, and Sequestration (CCTS).” Economics of Energy & Environmental Policy 2: 115–23. ———. 2012c. “How a ‘Low Carbon’ Innovation Can Fail -Tales from a ‘Lost Decade’ for Carbon Capture, Transport, and Sequestration (CCTS).” Economics of Energy & Environmental Policy 1 (2): 115–23. doi:10.5547/2160-5890.1.2.8. Hirschhausen, Christian von, Claudia Kemfert, Friedrich Kunz, and Roman Mendelevitch. 2013. “European Electricity Generation Post-2020: Renewable Energy Not To Be Underestimated.” 3. DIW Economic Bulletin. DIW Berlin. Ho, Minh T, Guy W Allinson, and Dianne E Wiley. 2011. “Comparison of MEA Capture Cost for Low CO2 Emissions Sources in Australia” International Journal of Greenhouse Gas Control (5 Issue 1): 49–60. Hobbs, B. F, and F. A.M Rijkers. 2004. “Strategic Generation with Conjectured Transmission Price Responses in a Mixed Transmission Pricing System-Part I: Formulation.” Power Systems, IEEE Transactions on 19 (2): 707–717. Hoel, Michael. 2012. “Carbon Taxes and the Green Paradox.” In Climate Change and Common SenseEssays in Honour of Tom Schelling, edited by Robert W. Hahn and Alistair Ulph. Oxford University Press. http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199692873.001.0001/acprof- 9780199692873. ———. 2013. “Supply Side Climate Policy and the Green Paradox.” http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2214897.
169 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Holz, Franziska, Clemens Haftendorn, Roman Mendelevitch, and Christian von Hirschhausen. 2015. “The COALMOD-World Model: Coal Markets until 2030.” In The Global Coal Market - Supplying the Major Fuel for Emerging Economies, edited by Richard K. Morse and Mark C. Thurber. Cambridge, UK: Cambridge University Press. ———. 2016. “DIW Berlin: A Model of the International Steam Coal Market (COALMOD-World).” DIW Data Documentation 85. Berlin: DIW Berlin. http://www.diw.de/documents/publikationen/73/diw_01.c.546364.de/diw_datadoc_2016- 085.pdf. Holz, Franziska, and Christian von Hirschhausen. 2013. “The Infrastructure Implications of the Energy Transformation in Europe until 2030 – Lessons from the EMF28 Modeling Exercise.” Climate Change Economics 4 (1): 1340006-1-1340006–26. doi:http://dx.doi.org/10.1142/S201000781340006X. Hotelling, Harold. 1931. “The Economics of Exhaustible Resources.” Journal of Political Economy 39 (2): 137–75. doi:10.1086/254195. Houde, Sébastien, and Anna C. Spurlock. 2016. “Minimum Energy Efficiency Standards for Appliances: Old and New Economic Rationales.” Economics of Energy & Environmental Policy 5 (2). doi:10.5547/2160-5890.5.2.shou. Houses of Parliament. 2012. “Low Carbon Technologies for Energy-Intensive Industries.” 403. London, UK: Parliamentary Office of Science & Technology. Hove, Sybille van den, Marc Le Menestrel, and Henri-Claude de Bettignies. 2002. “The Oil Industry and Climate Change: Strategies and Ethical Dilemmas.” Climate Policy 2 (1): 3–18. doi:10.1016/S1469-3062(02)00008-6. Hu, Jing, Wina Crijns-Graus, Long Lam, and Alyssa Gilbert. 2015. “Ex-Ante Evaluation of EU ETS during 2013–2030: EU-Internal Abatement.” Energy Policy 77 (February): 152–63. doi:10.1016/j.enpol.2014.11.023. Huntington, Hillard G, John P Weyant, and James L Sweeney. 1982. “Modeling for Insights, Not Numbers: The Experiences of the Energy Modeling Forum.” Omega 10 (5): 449–62. doi:10.1016/0305-0483(82)90002-0. Huppmann, Daniel, and Jonas Egerer. 2015. “National-Strategic Investment in European Power Transmission Capacity.” European Journal of Operational Research 247 (1): 191–203. doi:10.1016/j.ejor.2015.05.056. Huppmann, Daniel, and Ruud Egging. 2014. “Market Power, Fuel Substitution and Infrastructure – A Large-Scale Equilibrium Model of Global Energy Markets.” Energy 75: 483–500. doi:10.1016/j.energy.2014.08.004. Huppmann, Daniel, and Franziska Holz. 2012. “Crude Oil Market Power—A Shift in Recent Years?” The Energy Journal 33 (4): 1944–9089. doi:10.5547/01956574.33.4.1. Huppmann, Daniel, and Sauleh Siddiqui. 2015. “An Exact Solution Method for Binary Equilibrium Problems with Compensation and the Power Market Uplift Problem.” DIW Discussion Paper, DIW Berlin, 1475 (June). Idem, Raphael, Teeradet Supap, Huancong Shi, Don Gelowitz, Max Ball, Colin Campbell, and Paitoon Tontiwachwuthikul. 2015. “Practical Experience in Post-Combustion CO2 Capture Using Reactive Solvents in Large Pilot and Demonstration Plants.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 6–25. doi:10.1016/j.ijggc.2015.06.005. IEA. 1997. International Coal Trade: The Evolution of a Global Market. Paris France; Washington D.C.: OECD Washington Center [distributor].
———. 2005. “Building the Cost Curves for CO2 Storage: European Sector.” International Energy Agency. Paris, France.
———. 2009. “CO2 Capture and Storage - A Key Carbon Abatement Option.” International Energy Agency. Paris, France.
170 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
———. 2010. Energy Technology Perspectives 2010. Paris, France: International Energy Agency, OECD Publishing. https://www.iea.org/publications/freepublications/publication/etp2010.pdf. ———. 2011. Medium-Term Coal Market Report 2011. Medium-Term Coal Market Report. Paris, France: OECD Publishing. http://www.oecd-ilibrary.org/energy/medium-term-coal-market- report-2011_9789264167681-en. ———. 2012a. Energy Technology Perspectives 2012. Pathways to a Clean Energy System. Paris, France: International Energy Agency. ———. 2012b. Coal Information 2012. Coal Information. OECD Publishing. http://www.oecd- ilibrary.org/energy/coal-information-2012_coal-2012-en. ———. 2012c. Energy Statistics of Non-OECD Countries 2012. Energy Statistics of Non-OECD Countries. IEA. http://www.oecd-ilibrary.org/energy/energy-statistics-of-non-oecd-countries- 2012_energy_non-oecd-2012-en. ———. 2012d. World Energy Outlook 2012. Paris, France: International Energy Agency. http://www.oecd-ilibrary.org/energy/world-energy-outlook-2012_weo-2012-en. ———. 2012e. Medium-Term Coal Market Report 2012. Medium-Term Coal Market Report. Paris, France: OECD Publishing. http://www.oecd-ilibrary.org/energy/medium-term-coal-market- report-2012_9789264177963-en. ———. 2013a. “Coal Information.” International Energy Agency, Organisation for Economic Co- operation and Development. ———. 2014a. “Energy, Climate Change and Environment: 2014 Insights.” Energy, Climate Change and Environment. Paris, France: International Energy Agency, Organisation for Economic Co- operation and Development. http://www.oecd-ilibrary.org/energy/energy-climate-change-and- environment_9789264220744-en. ———. 2014b. Coal Information 2014. Coal Information. IEA. http://www.oecd-ilibrary.org/energy/coal- information-2014_coal-2014-en. ———. 2014c. World Energy Outlook 2014. World Energy Outlook. Paris, France: International Energy Agency, OECD Publishing. http://www.oecd-ilibrary.org/energy/world-energy-outlook- 2014_weo-2014-en. ———. 2015a. World Energy Outlook 2015. World Energy Outlook. Paris, France: International Energy Agency, OECD Publishing. http://www.oecd- ilibrary.org/docserver/download/6115271e.pdf?expires=1448556316&id=id&accname=ocid19 5153&checksum=B307C9BC7DFBD042C599EF4D6BE236CF. ———. 2015b. “World Energy Outlook Special Report 2015: Energy and Climate Change.” World Energy Outlook. http://www.worldenergyoutlook.org/energyclimate/#d.en.143801. ———. 2015c. Coal Information 2015. Coal Information. IEA. http://www.oecd-ilibrary.org/energy/coal- information-2015_coal-2015-en. ———. 2016. “IEA Energy Subsidies Database.” International Energy Agency, World Energy Outlook. http://www.worldenergyoutlook.org/media/weowebsite/2015/Subsidies20122014.xlsx.
IEA GHG, and ZEP. 2011. “The Costs of CO2 Storage - Post Demonstration CCS in the EU.” Brussels, Belgium: European Technology Platform for Zero Emission Power Plants (ZEP). IEA, International Energy Agency. 2013b. Coal Information 2013. OECD/IEA. IEA, NEA, and OECD. 2015. Projected Costs of Generating Electricity 2015. Paris, France: OECD Publishing. IEA, and UNIDO. 2011. “Technology Roadmap Carbon Capture and Storage in Industial Applications.” Paris, France: IEA Publications.
IEAGHG, and ZEP. 2011. “The Costs of CO2 Storage – Post Demonstration CCS in the EU.” IEA Greenhouse Gas R&D Programme (IEAGHG) and European Technology Platform for Zero Emission Power Plants (ZEP). Brussels, Belgium.
171 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
IEA/OECD. 2014. Medium-Term Coal Market Report 2014 - Market Analysis and Forecasts to 2019. Medium-Term Coal Market Report. OECD Publishing. http://www.oecd- ilibrary.org/energy/medium-term-coal-market-report-2014_mtrcoal-2014-en. IETA. 2012. “Briefing on the EU’s Emissions Trading Scheme.” Geneva, Switzerland: International Emissions Trading Association. IMF. 2015. “Counting the Cost of Energy Subsidies.” International Monetary Fund. 17. http://www.imf.org/external/np/fad/subsidies/data/codata.xlsx. Indian Bureau of Mines. 2012. “Indian Minerals Yearbook 2011.” Nagpur: Government of India. Ministry of Mines, Indian Bureau of Mines. Indiastat. 2012. “Mines and Minerals Statistics.” Available from http://www.indiastat.com. Accessed September 19, 2012. IPCC. 2005. IPCC Special Report on Carbon Dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press. https://www.ipcc.ch/pdf/special-reports/srccs/srccs_wholereport.pdf. ———. 2014. “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.” Geneva, Switzerland: IPCC. Jacobsen, Henriette. 2014. “Denmark Wants to Be Coal-Free by 2025.” EurActiv. November. http://www.euractiv.com/section/sustainable-dev/news/denmark-wants-to-be-coal-free-by- 2025/. Jansen, Daniel, Matteo Gazzani, Giampaolo Manzolini, Eric van Dijk, and Michiel Carbo. 2015. “Pre- Combustion CO2 Capture.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 167–87. doi:10.1016/j.ijggc.2015.05.028. Jenkins, Charles, Andy Chadwick, and Susan D. Hovorka. 2015. “The State of the Art in Monitoring and verification—Ten Years on.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 312– 49. doi:10.1016/j.ijggc.2015.05.009. Jevons, William Stanley. 1865. The Coal Question- An Inquiry Concerning the Progress of the Nation, and the Probable Exhaustion of Our Coal-Mines. 2nd ed. London, UK: Macmillan and Co. Johansson, Daniel J.A., Christian Azar, Kristian Lindgren, and Tobias A. Persson. 2009. “OPEC Strategies and Oil Rent in a Climate Conscious World.” The Energy Journal 30 (3): 23–50. Johnson, Nils, Volker Krey, David L. McCollum, Shilpa Rao, Keywan Riahi, and Joeri Rogelj. 2015. “Stranded on a Low-Carbon Planet: Implications of Climate Policy for the Phase-out of Coal- Based Power Plants.” Technological Forecasting and Social Change 90, Part A (January): 89– 102. doi:10.1016/j.techfore.2014.02.028. Johnson, Nils, Manfred Strubegger, Madeleine McPherson, Simon C. Parkinson, Volker Krey, and Patrick Sullivan. 2016. “A Reduced-Form Approach for Representing the Impacts of Wind and Solar PV Deployment on the Structure and Operation of the Electricity System.” Energy Economics, July. doi:10.1016/j.eneco.2016.07.010. Jones, D. G., S. E. Beaubien, J. C. Blackford, E. M. Foekema, J. Lions, C. De Vittor, J. M. West, S. Widdicombe, C. Hauton, and A. M. Queirós. 2015. “Developments since 2005 in Understanding Potential Environmental Impacts of CO2 Leakage from Geological Storage.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 350–77. doi:10.1016/j.ijggc.2015.05.032. Jones, Dave, and Kathrin Gutmann. 2015. “End of an Era: Why Every European Country Needs a Coal Phase-Out Plan.” London, UK: Greenpeace UK, Climate Action Network (CAN) Europe.
172 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
JORC. 2012. “Australasian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves - 2012 Edition.” Carlton South, Australia: The Joint Ore Reserves Committee of The Australasian Institute of Mining and Metallurgy, Australian Institute of Geoscientists and Minerals Council of Australia. JRC. 2011. “2011 Technology Map of the European Strategic Energy Technology Plan (SET-Plan) - Technology Descriptions.” JRC Scientific and Technical Reports. Petten, The Netherlands: European Commission, Joint Research Centre, Institute for Energy and Transport. https://setis.ec.europa.eu/system/files/Technology_Map_2011.pdf. Kalkuhl, Matthias, and Robert J. Brecha. 2013. “The Carbon Rent Economics of Climate Policy.” Energy Economics 39 (September): 89–99. doi:10.1016/j.eneco.2013.04.008. Kannan, Deepak, Arusha Das, and Mia Corazon Aureus. 2015. “Indonesia Plans New Tax on Thermal Coal Exports from August 8.” Coal | Platts News Article & Story. July 6. http://www.platts.com/latest-news/coal/singapore/indonesia-plans-new-tax-on-thermal-coal- exports-27577871. Kanudia, Amit, Niels Berghout, Dulce Boavida, Machteld van den Broek, Helena Cabal, Júlio Carneiro, Patrícia Fortes, et al. 2013. “CCS Infrastructure Development Scenarios for the Integrated Iberian Peninsula and Morocco Energy System.” Energy Procedia, GHGT-11 Proceedings of the 11th International Conference on Greenhouse Gas Control Technologies, 18-22 November 2012, Kyoto, Japan, 37: 2645–56. doi:10.1016/j.egypro.2013.06.149. Kazmierczak, Tomasz, Ruut Brandsma, Filip Neele, and Chris Hendriks. 2008. “Algorithm to Create a CCS Low-Cost Pipeline Network.” Energy Procedia 1 (1): 1617–23. Kemfert, Claudia, Jochen Diekmann, and Hans-Joachim Ziesing. 2004. “Emissions Trading in Europe: Effective Tool or Flight of Fancy?” Intereconomics 39 (3): 119–21.
Kemp, Alexander G., and Sola Kasim. 2013. “The Economics of CO2-EOR Cluster Developments in the UK Central North Sea.” Energy Policy 62 (November): 1344–55. doi:10.1016/j.enpol.2013.07.047. Kemper, Jasmin. 2015. “Biomass and Carbon Dioxide Capture and Storage: A Review.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 401–30. doi:10.1016/j.ijggc.2015.06.012. Kjärstad, Jan, Joris Morbee, Mikael Odenberger, Filip Johnsson, and Evangelos Tzimas. 2013. “Modelling Large-Scale CCS Development in Europe Linking Techno-Economic Modelling to Transport Infrastructure.” Energy Procedia, GHGT-11, 37: 2941–48. doi:10.1016/j.egypro.2013.06.180.
Klokk, O., P. Schreiner, A. Pagès-Bernaus, and A. Tomasgard. 2010. “Optimizing a CO2 Value Chain for the Norwegian Continental Shelf.” Energy Policy 38: 6604–14. Knoope, M. M. J., A. Ramírez, and A. P. C. Faaij. 2015. “The Influence of Uncertainty in the Development of a CO2 Infrastructure Network.” Applied Energy 158 (November): 332–47. doi:10.1016/j.apenergy.2015.08.024. Knopf, Brigitte, Yen-Heng Henry Chen, Enrica De Cian, Hannah Förster, Amit Kanudia, Ioanna Karkatsouli, Ilkka Keppo, Tiina Koljonen, Katja Schumacher, and Detlef P. Van Vuuren. 2013. “Beyond 2020 - Strategies and Costs for Transforming the European Energy System.” Climate Change Economics 04 (supp01): 1340001. doi:10.1142/S2010007813400010. Kobos, Peter H., Leonard A. Malczynski, David J. Borns, and Brian J. McPherson. 2007. “The ‘String of Pearls’: The Integrated Assessment Cost and Source-Sink Model.” Conference Proceedings of the 6th Annual Carbon Capture & Sequestration Conference. Pittsburgh, PA, USA. Kojima, Masami, and Doug Koplow. 2015. “Fossil Fuel Subsidies: Approaches and Valuation.” Working Paper. World Bank Group, Washington, DC. https://openknowledge.worldbank.org/handle/10986/21659. Kolstad, Charles D., and David S. Abbey. 1984. “The Effect of Market Conduct on International Steam Coal Trade.” European Economic Review 24 (1): 39–59. doi:10.1016/0014-2921(84)90012-6.
173 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Kolstad, Charles D., Kevin Urama, John Broome, Annegrete Bruvoll, M. Cariño-Olvera, Don Fullerton, Christian Gollier, et al. 2014. “Social, Economic and Ethical Concepts and Methods.” In Climate Change 2014: Mitigation of Climate Change, 173–248. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press. http://lup.lub.lu.se/record/5104578. Kolstad, Charles D., and Jr. Wolak Frank A. 1985. “Strategy and Market Structure in Western Coal Taxation.” The Review of Economics and Statistics 67 (2): 239–49. Koplow, D. 2015. “Subsidies to Energy Industries.” In Reference Module in Earth Systems and Environmental Sciences. Elsevier. http://www.sciencedirect.com/science/article/pii/B9780124095489092691. Kossoy, Alexandre, Grzegorz Peszko, Klaus Oppermann, Nicolai Prytz, Noemie Klein, Kornelis Blok, Long Khanh Lam, Lindee Wong, and Bram Borkent. 2015a. “State and Trends of Carbon Pricing 2015.” 99533. The World Bank. http://documents.worldbank.org/curated/en/2015/09/25053834/state-trends-carbon-pricing- 2015. ———. 2015b. “State and Trends of Carbon Pricing 2015.” 99533. Washington DC, USA: The World Bank. http://documents.worldbank.org/curated/en/2015/09/25053834/state-trends-carbon- pricing-2015. Krevor, Samuel, Martin J. Blunt, Sally M. Benson, Christopher H. Pentland, Catriona Reynolds, Ali Al- Menhali, and Ben Niu. 2015. “Capillary Trapping for Geologic Carbon Dioxide Storage – From Pore Scale Physics to Field Scale Implications.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 221–37. doi:10.1016/j.ijggc.2015.04.006. Kuby, Michael J., Jeffrey M. Bielicki, and Richard S. Middleton. 2011. “Optimal Spatial Deployment of Carbon Dioxide Capture and Storage Given a Price on Carbon Dioxide,” Special Issue for ISOLDE XI, , 34:285. Kunith, Alexander, Roman Mendelevitch, and Dietmar Göhlich. 2016. “Market Power Rents and Climate Change Mitigation: A Rationale for Coal Taxes?” DIW Berlin, Discussion Paper 1577. Berlin, Germany: German Institute for Economic Research (DIW Berlin). http://www.diw.de/documents/publikationen/73/diw_01.c.534056.de/dp1577.pdf. Kunz, Friedrich, and A. Zerrahn. 2015. “Benefits of Coordinating Congestion Management in Electricity Transmission Networks: Theory and Application to Germany.” Utilities Policy Available online 21 October 2015: 1–12. doi:10.1016/j.jup.2015.09.009.
Kuramochi, T., A. Ramírez, W. Turkenburg, and André Faaij. 2012. “Comparative Assessment of CO2 Capture Technologies for Carbon-Intensive Industrial Processes.” Progress in Energy and Combustion Science 38 (1): 87–112. Lazarus, Michael, Peter Erickson, and Kevin Tempest. 2015. “Supply-Side Climate Policy: The Road Less Taken,” Stockholm Environment Institute, Working Paper, , no. SEI-WP-2015-13 (October): 1–24. Leader of the G7. 2015. “Leadersʼ Declaration G7 Summit, 7– 8 June 2015.” Schloss Elmau, Germany. Leaton, James, Nicola Ranger, Bob Ward, Luke Sussams, and Meg Brown. 2013. “Unburnable Carbon 2013: Wasted Capital and Stranded Assets.” Carbon Tracker Initiative and Grantham Research Institute. Lévêque, François. 2014. The Economics and Uncertainties of Nuclear Power. Cambridge, UK: Cambridge University Press. Liang, Zhiwu Henry, Wichitpan Rongwong, Helei Liu, Kaiyun Fu, Hongxia Gao, Fan Cao, Rui Zhang, et al. 2015. “Recent Progress and New Developments in Post-Combustion Carbon-Capture Technology with Amine Based Solvents.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 26–54. doi:10.1016/j.ijggc.2015.06.017.
174 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Lin, Boqiang, and Xiaoling Ouyang. 2014. “A Revisit of Fossil-Fuel Subsidies in China: Challenges and Opportunities for Energy Price Reform.” Energy Conversion and Management 82 (June): 124– 34. doi:10.1016/j.enconman.2014.03.030. Lise, Wietze, Benjamin F. Hobbs, and Frits van Oostvoorn. 2008. “Natural Gas Corridors between the EU and Its Main Suppliers: Simulation Results with the Dynamic GASTALE Model.” Energy Policy 36 (6): 1890–1906. doi:10.1016/j.enpol.2008.01.042. Liski, Matti, and Olli Tahvonen. 2004. “Can Carbon Tax Eat OPEC’s Rents?” Journal of Environmental Economics and Management 47 (1): 1–12. doi:10.1016/S0095-0696(03)00052-4. Liu, Gang. 2004. “Estimating Energy Demand Elasticities for OECD Countries. A Dynamic Panel Data Approach.” Discussion Paper. Statistics Norway, Research Department. http://econpapers.repec.org/paper/ssbdispap/373.htm. Lontoh, Lucky, and Christopher Beaton. 2015. “G20 Subsidies to Oil, Gas and Coal Production: Indonesia.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/9964.pdf. Lucarelli, Bart. 2015. “Government as Creator and Destroyer.” In The Global Coal Market: Supplying the Major Fuel for Emerging Economies, edited by Mark C. Thurber and Richard K. Morse. Cambridge, UK: Cambridge University Press. Lupion, Monica, and Howard J. Herzog. 2013. “NER300: Lessons Learnt in Attempting to Secure CCS Projects in Europe.” International Journal of Greenhouse Gas Control 19 (November): 19–25. doi:10.1016/j.ijggc.2013.08.009. Makhijani, Shakuntala, and Alex Doukas. 2015. “G20 Subsidies to Oil, Gas and Coal Production: Australia.” Background Paper for the Report Empty Promises: G20 Subsidies to Oil, Gas and Coal Production. Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/publications/10071-g20-subsidies-oil-gas-coal-production-australia. Marshall, Jonathan Paul. 2016. “Disordering Fantasies of Coal and Technology: Carbon Capture and Storage in Australia.” Energy Policy. Accessed August 12. doi:10.1016/j.enpol.2016.05.044. Martin, Aran. 2014. “Commodity Exports and Transboundary Atmospheric Impacts: Regulating Coal in an Era of Climate Change.” Environmental Politics 23 (4): 590–609. doi:10.1080/09644016.2014.891788. Martin, Pamela L. 2014. “Ecuador’s Yasuní-ITT Initiative: Why Did It Fail?’.” International Development Policy| Revue Internationale de Politique de Développement 6: 2014. Martínez Arranz, Alfonso. 2015. “Carbon Capture and Storage: Frames and Blind Spots.” Energy Policy 82: 249–259. Massol, Olivier, Stéphane Tchung-Ming, and Albert Banal-Estañol. 2015. “Joining the CCS Club! The Economics of CO2 Pipeline Projects.” European Journal of Operational Research 247 (1): 259–75. doi:10.1016/j.ejor.2015.05.034. McCracken, Ross. 2015. “King Coal Faces the End of Its Reign.” The Barrel. March 25. http://blogs.platts.com/2015/03/25/king-coal-end/. McGlade, Christophe, and Paul Ekins. 2015a. “The Geographical Distribution of Fossil Fuels Unused When Limiting Global Warming to 2 °C.” Nature 517 (7533): 187–90. doi:10.1038/nature14016. ———. 2015b. “The Geographical Distribution of Fossil Fuels Unused When Limiting Global Warming to 2 °C.” Nature 517 (7533): 187–90. doi:10.1038/nature14016. McPherson, B., R. Allis, B. Biediger, J. Brown, and G. Cappa. 2009. “Southwest Regional Partnership on Carbon Sequestration.” Revised Semiannual Report. Socorro, New Mexico, USA: New Mexico Institute of Mining and Technology. Meinshausen, Malte, Nicolai Meinshausen, William Hare, Sarah C. B. Raper, Katja Frieler, Reto Knutti, David J. Frame, and Myles R. Allen. 2009. “Greenhouse-Gas Emission Targets for Limiting Global Warming to 2°C.” Nature 458 (7242): 1158–62. doi:10.1038/nature08017.
175 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Mendelevitch, Roman. 2014. “The Role of CO₂-EOR for the Development of a CCTS Infrastructure in the North Sea Region: A Techno-Economic Model and Applications.” International Journal of Greenhouse Gas Control 20 (January): 132–59. doi:10.1016/j.ijggc.2013.11.007. ———. 2016a. “Coal Markets and Carbon Capture - Model Development and Climate Policy Applications.” Doctoral Thesis. Berlin, Germany. ———. 2016b. “Testing Supply-Side Climate Policies for the Global Steam Coal Market – Can They Curb Coal Consumption?” DIW Berlin, Discussion Paper 1604. Berlin, Germany: German Institute for Economic Research (DIW Berlin). Mendelevitch, Roman, and Thanh Thien Dang. 2016. “Nuclear Power and the Uranium Market: Are Reserves and Resources Sufficient?” 98. DIW Roundup. Berlin, Germany: German Institute for Economic Research (DIW Berlin). https://www.diw.de/documents/publikationen/73/diw_01.c.536014.de/diw_roundup_98_en.pdf.
Mendelevitch, Roman, J. Herold, P.-Y. Oei, and A. Tissen. 2010. “CO2 Highways for Europe Modelling a Carbon Capture, Transport and Storage Infrastructure for Europe.” CEPS Policy Brief. Brussels, Belgium. http://www.ceps.eu/book/CO2-highways-europe-modelling-carbon-capture- transport-and-storage-infrastructure-europe. Mendelevitch, Roman, and Pao-Yu Oei. 2015. “Consequences of the UK Energy Market Reform on the Development of Carbon Capture, Transport, and Storage.” In , 1–6. Lisbon, Portugal. doi:10.1109/EEM.2015.7216683. Merrill, Laura. 2014. “Fossil-Fuel Subsidy Reform - Mitigating Emissions through Getting the Price Right.” COP 20 Workshop presented at the Global Subsidies Initiative, International Institute for Sustainable Development, Lima, Peru, December 5. Merrill, Laura, Melissa Harris, Liesbeth Casier, and Andrea Bassi. 2015. “Fossil-Fuel Subsidies and Climate Change.” Nordic Council of Ministers. http://urn.kb.se/resolve?urn=urn:nbn:se:norden:org:diva-3744. Metayer, Matthieu, Christian Breyer, and Hans-Josef Fell. 2015. “The Projections for the Future and Quality in the Past of the World Energy Outlook for Solar PV and Other Renewable Energy Technologies.” Energy Watch Group. Middleton, R., H. Herzog, G. Keating, M. Kuby, and X. Liao. 2007. “Optimization for Geologic Carbon Sequestration and Carbon Credit Pricing.” Conference Proceedings of the 6th Annual Carbon Capture & Sequestration Conference. Pittsburgh, PA, USA. Middleton, Richard S., and Jeffrey M. Bielicki. 2009. “A Scalable Infrastructure Model for Carbon Capture and Storage: SimCCS.” Energy Policy 37 (3): 1052–60. doi:DOI: 10.1016/j.enpol.2008.09.049.
Middleton, Richard S., and Jordan K. Eccles. 2013. “The Complex Future of CO2 Capture and Storage: Variable Electricity Generation and Fossil Fuel Power.” Applied Energy 108 (August): 66–73. doi:10.1016/j.apenergy.2013.02.065.
Milligan, Ben. 2014. “Planning for Offshore CO2 Storage: Law and Policy in the United Kingdom.” Marine Policy 48 (September): 162–71. doi:10.1016/j.marpol.2014.03.029. MIT. 2015. “Energy and Climate Outlook 2015.” Cambridge, MA., USA: MIT Joint Program on the Science and Policy of Global Change. http://globalchange.mit.edu/files/2015%20Energy%20%26%20Climate%20Outlook.pdf. ———. 2016. “Carbon Capture and Sequestration Technologies @ MIT.” Carbon Capture and Sequestration Project Database. http://sequestration.mit.edu/tools/projects/index_capture.html. Mooney, Chris, and Steven Mufson. 2016. “How Coal Titan Peabody, the World’s Largest, Fell into Bankruptcy.” The Washington Post, April 13. https://www.washingtonpost.com/news/energy- environment/wp/2016/04/13/coal-titan-peabody-energy-files-for-bankruptcy/. Morbee, J., J. Serpa, and E. Tzimas. 2012. “Optimised Deployment of a European CO₂ Transport Network.” International Journal of Greenhouse Gas Control 7 (March): 48–61. doi:10.1016/j.ijggc.2011.11.011.
176 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Morse, Richard K., and Gang He. 2015. “The World’s Greatest Coal Arbitrage.” In The Global Coal Market: Supplying the Major Fuel for Emerging Economies, edited by Mark C. Thurber and Richard K. Morse. Cambridge, UK: Cambridge University Press. Mou, Dunguo, and Zhi Li. 2012. “A Spatial Analysis of China’s Coal Flow.” Energy Policy 48 (September): 358–68. doi:10.1016/j.enpol.2012.05.034. NBSC. 2007. “China Energy Statistical Yearbook 2007.” China Statistics Press. Beijing, China: National Bureau of Statistics of China. ———. 2015. “Statistical Communiqué of the People’s Republic of China on the 2014 National Economic and Social Development.” February. http://www.stats.gov.cn/english/PressRelease/201502/t20150228_687439.html. ———. 2016. “China Statistical Yearbook-2014.” National Bureau of Statistics China. http://www.stats.gov.cn/tjsj/ndsj/2014/indexee.htm. Neele, Filip, Chris Hendriks, and Ruut Brandsma. 2009. “Geocapacity: Economic Feasibility of CCS in Networked Systems.” Greenhouse Gas Control Technologies 9, Proceedings of the 9th International Conference on Greenhouse Gas Control Technologies (GHGT-9), 16-20 November 2008, Washington DC, USA 1 (1): 4217–4224. Neuhoff, K., W. Acworth, R. Betz, D. Burtraw, J. Cludius, H. Fell, C. Hepburn, et al. 2015. “Is a Market Stability Reserve Likely to Improve the Functioning of the EU ETS?: Evidence from a Model Comparison Exercise.” VIII. Climate Strategies. London, UK. http://www.diw.de/sixcms/detail.php?id=diw_01.c.497154.de. Nordhaus, William D. 1991. “To Slow or Not to Slow: The Economics of The Greenhouse Effect.” The Economic Journal 101 (407): 920–37. doi:10.2307/2233864. ———. 2013. “Chapter 16 - Integrated Economic and Climate Modeling.” In Handbook of Computable General Equilibrium Modeling, edited by Peter Jorgenson, B. Dixon, and W. Dale, 1:1069– 1131. Handbook of Computable General Equilibrium Modeling SET, Vols. 1A and 1B. Elsevier. http://www.sciencedirect.com/science/article/pii/B978044459568300016X. NSW Department of Industry. 2015. “Thermal Coal: Opportunities in New South Wales, Australia.” NSW Department of Industry, Division of Resources & Energy. http://www.resourcesandenergy.nsw.gov.au/__data/assets/pdf_file/0008/581606/thermal- coal.pdf. NSWDPI. 2009. “Summary of NSW Coal Statistics.” New South Wales Department of Primary Industries. Available from http://www.resources.nsw.gov.au/resources/coal. OCI. 2015. “Indonesia: Public Finance/Subsidies for Coal.” Oil Change International. http://priceofoil.org/content/uploads/2014/06/OCI-Indonesia-Coal-Subsidies-Summary- February-2014.pdf. ODI. 2015a. “G20 Subsidies to Oil Gas and Coal Production: China Data Sheet.” ODI Shaping Policy for Development. November 2. https://www.odi.org/sites/odi.org.uk/files/odi- assets/publications-opinion-files/9987.xlsx. ———. 2015b. “G20 Subsidies to Oil Gas and Coal Production: India Data Sheet.” ODI Shaping Policy for Development. November 11. https://www.odi.org/sites/odi.org.uk/files/odi- assets/publications-opinion-files/9960.xlsx. ———. 2015c. “G20 Subsidies to Oil Gas and Coal Production: United States Data Sheet.” ODI Shaping Policy for Development. November 11. https://www.odi.org/sites/odi.org.uk/files/odi- assets/publications-opinion-files/9980.xlsx. OECD. 2013a. “Climate and Carbon: Aligning Prices and Policies.” OECD Environment Policy Paper. Paris, France. http://www.oecd-ilibrary.org/environment-and-sustainable-development/climate- and-carbon_5k3z11hjg6r7-en. ———. 2013b. “Effective Carbon Prices.” OECD Publishing. http://www.oecd- ilibrary.org/environment/effective-carbon-prices_9789264196964-en. ———. 2015a. “Fossil Fuel Support.” Inventory of Estimates Budgetary Support and Tax Expenditures for Fossil Fuels 2015. September. http://www.oecd.org/site/tadffss/.
177 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
———. 2015b. OECD Companion to the Inventory of Support Measures for Fossil Fuels 2015. OECD Publishing. http://www.oecd-ilibrary.org/energy/oecd-companion-to-the-inventory-of-support- measures-for-fossil-fuels-2015_9789264239616-en. OECD, and IEA. 2008a. World Energy Outlook 2008. Paris, France: OECD Publishing; International Energy Agency. http://public.eblib.com/EBLPublic/PublicView.do?ptiID=407996. ———. 2008b. World Energy Outlook 2008. Paris, France: Organisation for Economic Co-operation and Development. http://public.eblib.com/EBLPublic/PublicView.do?ptiID=407996. OECD/IEA. 2015. “Medium-Term Coal Market Report 2015.” Paris, France: International Energy Agency. http://www.oecd-ilibrary.org/content/book/mtrcoal-2015-en. ———. 2016. “Glossary.” International Energy Agency, Secure, Sustainable, Together. http://www.iea.org/aboutus/glossary/#tabs-2. Oei, Pao-Yu. 2016. “Decarbonizing the European Electricity Sector.” doi:10.14279/depositonce-5065. Oei, Pao-Yu, Johannes Herold, and Roman Mendelevitch. 2014. “Modeling a Carbon Capture, Transport, and Storage Infrastructure for Europe.” Environmental Modeling & Assessment 19 (May): 515–31. doi:10.1007/s10666-014-9409-3.
Oei, Pao-Yu, Johannes Herold, and Andreas Tissen. 2011. “CO2 Speicherung in Deutschland: Eine Brückentechnologie als Klimalösung?” Zeitschrift für Energiewirtschaft 35 (4): 263–73. doi:10.1007/s12398-011-0061-1. Oei, Pao-Yu, Claudia Kemfert, Felix Reitz, and Christian von Hirschhausen. 2014a. “Braunkohleausstieg – Gestaltungsoptionen Im Rahmen Der Energiewende.” DIW Berlin, Politikberatung kompakt 84. Berlin, Germany: DIW Berlin — Deutsches Institut für Wirtschaftsforschung e. V. https://www.diw.de/documents/publikationen/73/diw_01.c.471589.de/diwkompakt_2014- 084.pdf. ———. 2014b. “Coal Power Endangers Climate Targets: Calls for Urgent Action.” 8. Economic Bulletin. Berlin, Germany: German Institute for Economic Research. Oei, Pao-Yu, and Roman Mendelevitch. 2013. “Development Scenarios for a CO₂ Infrastructure Network in Europe.” Resource Markets Working Paper WP-RM-36. Potsdam, Germany: Universität Potsdam. http://www.uni- potsdam.de/fileadmin/projects/wipo/Projekt_RM/WP_RM_36_Oei_Mendelevitch_2013_CO2_I nfrastructure_Europe.pdf. Oei, Pao-Yu, Roman Mendelevitch, Johannes Herold, Andreas Tissen, and Christian von Hirschhausen. 2010. “CO2 Autobahnen Für Europa?” Energiewirtschaftliche Tagesfragen, ETV Energieverlag GmbH, , no. 12/2010: 46–52. Ogarenko, Iuliia, Ken Bossong, Ivetta Gerasimchuk, and Sam Pickard. 2015. “G20 Subsidies to Oil, Gas and Coal Production: Russia.” Oil Change International (OCI) and the Overseas Development Institute (ODI). https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications- opinion-files/9969.pdf.
Öko-Institut. 2012. “Potenziale Und Chancen Der Technologie Zur CO2–Abtrennung Und Ablagerung (CCS) Für Industrielle Prozessemissionen.” Kurzstudie für die Umweltstiftung WWF Deutschland. Berlin, Germany: Öko - Institut e.V. Osborne, D., ed. 2013. The Coal Handbook: Towards Cleaner Production. Volume 2. Woodhead Publishing Series in Energy, no. 51. Philadelphia, USA: Woodhead Pub. Osborne, George. 2014. “Chancellor George Osborne’s Budget 2014 Speech.” 19th March 2014. London, UK. Ostrom, Elinor. 2010. “Beyond Markets and States: Polycentric Governance of Complex Economic Systems.” The American Economic Review 100 (3): 641–72. doi:10.2307/27871226. Paltsev, Sergey V. 2001. “The Kyoto Protocol: Regional and Sectoral Contributions to the Carbon Leakage.” Energy Journal 22 (4): 53–79.
178 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Parliament of Canada. 2012. Reduction of Carbon Dioxide Emissions from Coal-Fired Generation of Electricity Regulations. SOR/2012-167. Ottawa, Ontario, Canada. http://laws.justice.gc.ca/eng/regulations/sor-2012-167/FullText.html. Paulus, Moritz, and Johannes Trüby. 2011. “Coal Lumps vs. Electrons: How Do Chinese Bulk Energy Transport Decisions Affect the Global Steam Coal Market?” Energy Economics 33 (6): 1127– 37. doi:10.1016/j.eneco.2011.02.006. Perman, Roger, Yue Ma, Michael Common, David Maddison, and James Mcgilvray. 2012. Natural Resource and Environmental Economics. 4 edition. Harlow, Essex ; New York: Prentice Hall. Persson, Tobias A., Christian Azar, Daniel J.A. Johansson, and Kristian Lindgren. 2007. “Major Oil Exporters May Profit rather than Lose, in a Carbon-Constrained World.” Energy Policy 35 (12): 6346–53. doi:10.1016/j.enpol.2007.06.027. Pfenninger, Stefan, Adam Hawkes, and James Keirstead. 2014. “Energy Systems Modeling for Twenty-First Century Energy Challenges.” Renewable and Sustainable Energy Reviews 33: 74–86. doi:10.1016/j.rser.2014.02.003. Pindyck, Robert S. 2015. “The Use and Misuse of Models for Climate Policy.” 21097. NBER Working Paper. Cambridge, MA: National Bureau of Economic Research. http://www.nber.org/papers/w21097.pdf. Platts. 2009. “Methodology and Specification Guide – Coal.” Available from www.platts.com. ———. 2011. “Power System Database Europe.” London, UK: Platts, McGraw Hill Financial. Pollitt, Michael G., and Aoife Brophy Haney. 2013. “Dismantling a Competitive Electricity Sector: The U.K.’s Electricity Market Reform.” The Electricity Journal 26 (10): 8–15. doi:10.1016/j.tej.2013.11.001. PwC. 2015. “Mining in Indonesia 2015.” Jakarta, Indonesia: PricewaterhouseCoopers International Limited, Indonesia Office. QLDDME. 2009. “Coal Statistics.” Queensland Department of Mines and Energy. Available from http://mines.industry.qld.gov.au/. Queensland Government. 2016. “Production of Saleable Coal by Individual Mines.” Coal Industry Review Statistical Tables - Data | Queensland Government. May 18. https://data.qld.gov.au/dataset/coal-industry-review-statistical-tables/resource/1b7fb643-c880- 42bf-940b-fc3c582d239d. Rademacher, Maggi. 2008. “Development and Perspectives on Supply and Demand in the Global Hard Coal Market.” Zeitschrift Für Energiewirtschaft 32 (2): 67–87. doi:10.1007/s12398-008- 0010-9.
Radoslaw, Tarkowski, Uliasz-Misiak Barbara, and Wójcicki Adam. 2009. “CO2 Storage Capacity of Deep Aquifers and Hydrocarbon Fields in Poland – EU GeoCapacity Project Results.” Energy Procedia 1 (1): 2671–77. Rajagopal, Krishnadas. 2014. “Supreme Court Quashes Allocation of 214 Coal Blocks.” The Hindu, September 24. http://www.thehindu.com/news/national/supreme-court-quashes-allocation-of- all-but-four-of-218-coal-blocks/article6441855.ece. RECCS. 2010. “RECCS plus: Comparison of Renewable Energy Technologies with Carbon Dioxide Capture and Storage.” Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (BMU), Berlin, Germany. Reiner, David M. 2016. “Learning through a Portfolio of Carbon Capture and Storage Demonstration Projects.” Nature Energy 1 (January): 15011. doi:10.1038/nenergy.2015.11. Richter, Philipp M. 2015. “Trade and Climate Policies: The Role of Market Structure, Strategic Behaviour, and the Supply Side of Carbon.” Berlin, Germany: Technische Universität Berlin. https://opus4.kobv.de/opus4-tuberlin/files/6914/richter_philipp.pdf. Richter, Philipp M., Roman Mendelevitch, and Frank Jotzo. 2015. “Market Power Rents and Climate Change Mitigation: A Rationale for Coal Taxes?” DIW Berlin, Discussion Paper 1471. Berlin, Germany: German Institute for Economic Research (DIW Berlin). http://www.diw.de/sixcms/detail.php?id=diw_01.c.502680.de.
179 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Rioux, Bertrand, Philipp Galkin, Frederic Murphy, and Axel Pierru. 2015. “Economic Impacts of Debottlenecking Congestion in the Chinese Coal Supply Chain.” KS-1523-DP017A. KAPSARC Discussion Paper. Riyadh, Saudi Arabia: King Abdullah Petroleum Studies and Research Center (KAPSARC). https://www.kapsarc.org/wp-content/uploads/2015/10/KS- 1523-DP017A-Economic-Impacts-of-Debottlenecking-Congestion-in-the-Chinese-Coal- Supply-Chain.pdf. Ritschel, Wolfgang, and Hans-Wilhelm Schiffer. 2007. “World Market for Hard Coal, 2007 Edition.” Essen, Germany: RWE Power AG. Rogner, Hans-Holger, Roberto F. Aguilera, Christina Archer, Ruggero Bertani, S. C. Bhattacharya, Maurice B. Dusseault, Luc Gagnon, et al. 2012. “Energy Resources and Potentials.” In Global Energy Assessment - Toward a Sustainable Future, 423–512. Cambridge University Press, Cambridge, UK and New York, NY, USA and the International Institute for Applied Systems Analysis, Laxenburg, Austria. www.globalenergyassessment.org. Rosenow, Jan, Cor Leguijt, Zsuzsanna Pato, Nick Eyre, and Tina Fawcet. 2016. “An Ex-Ante Evaluation of the EU Energy Efficiency Directive - Article 7.” Economics of Energy & Environmental Policy 5 (2). doi:10.5547/2160-5890.5.2.jros.
Rubin, Edward S., John E. Davison, and Howard J. Herzog. 2015. “The Cost of CO2 Capture and Storage.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 378–400. doi:10.1016/j.ijggc.2015.05.018. Rubin, Edward S., M. Antes S. Yeh, M. Berkenpas, and J. Davison. 2007. “Use of Experience Curves to Estimate the Future Cost of Power Plants with CO2 Capture.” International Journal of Greenhouse Gas Control 1 (2): 188–197. SACRM. 2011. “Overview of the South African Coal Value Chain - Prepared as a Basis for the Development of the South African Coal Roadmap.” South African Coal Roadmap. http://www.fossilfuel.co.za/initiatives/2013/SACRM-Value-Chain-Overview.pdf. Sanzillo, Tom. 2015. “Indonesia’s Coal-Trafficking Epidemic Requires an Impartial Inquiry.” Institute for Energy Economics & Financial Analysis. June 15. http://ieefa.org/indonesias-coal-trafficking- epidemic-requires-an-impartial-inquiry/. Schernikau, Lars. 2010. Economics of the International Coal Trade: The Renaissance of Steam Coal. Springer Science & Business Media. Schneider, Mycle, Antony Froggatt, Julie Hazemann, Ian Fairlie, Tadahiro Katsuta, Fulcieri Maltini, and M.V. Ramana. 2016. “World Nuclear Industry Status Report 2016.” Paris. http://www.worldnuclearreport.org/IMG/pdf/20160713MSC-WNISR2016V2-HR.pdf. Schröder, Andreas, Friedrich Kunz, Jan Meiß, Roman Mendelevitch, and Christian von Hirschhausen. 2013. “Current and Prospective Costs of Electricity Generation until 2050.” DIW Berlin, Data Documentation 68. Berlin, Germany: DIW Berlin. http://www.diw.de/documents/publikationen/73/diw_01.c.424566.de/diw_datadoc_2013- 068.pdf.
Schumann, Diana, Elisabeth Duetschke, and Katja Pietzner. 2014. “Public Perception of CO2 Offshore Storage in Germany: Regional Differences and Determinants.” 12th International Conference on Greenhouse Gas Control Technologies, GHGT-12 63: 7096–7112. doi:10.1016/j.egypro.2014.11.744. Schwanitz, Valeria Jana, Franziska Piontek, Christoph Bertram, and Gunnar Luderer. 2014. “Long- Term Climate Policy Implications of Phasing out Fossil Fuel Subsidies.” Energy Policy 67 (April): 882–94. doi:10.1016/j.enpol.2013.12.015. Scott, Vivian, Stuart Gilfillan, Nils Markusson, Hannah Chalmers, and R. Stuart Haszeldine. 2013. “Last Chance for Carbon Capture and Storage.” Nature Climate Change 3 (2): 105–11. doi:10.1038/nclimate1695.
Scottish Centre for Carbon Storage. 2009. Opportunities for CO2 Storage Around Scotland: An Integrated Strategic Research Study. Edited by Scottish Government. Edinburgh, Scotland:
180 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Scottish Centre for Carbon Storage. http://www.scotland.gov.uk/Publications/2009/04/28114540/0. Shapiro, Jeremy F., and David E. White. 1982. “A Hybrid Decomposition Method for Integrating Coal Supply and Demand Models.” Operations Research 30 (5): 887–906. doi:10.1287/opre.30.5.887. Siddiqui, Sauleh. 2011. Solving Two-Level Optimization Problems with Applications to Robust Design and Energy Markets. Siddiqui, Sauleh, and Steven A. Gabriel. 2013. “An SOS1-Based Approach for Solving MPECs with a Natural Gas Market Application.” Networks and Spatial Economics 13 (2): 205–27. doi:10.1007/s11067-012-9178-y. Sinn, Hans-Werner. 2008. “Public Policies against Global Warming: A Supply Side Approach.” International Tax and Public Finance 15 (4): 360–94. doi:10.1007/s10797-008-9082-z. ———. 2015. “Introductory Comment - The Green Paradox: A Supply-Side View of the Climate Problem.” Review of Environmental Economics and Policy 9 (2): 239–45. doi:10.1093/reep/rev011. Smil, Vaclav. 2000. “Energy in the Twentieth Century: Resources, Conversions, Costs, Uses, and Consequences.” Annual Review of Energy and the Environment 25 (1): 21–51. doi:10.1146/annurev.energy.25.1.21. Sobczyk, E., and P. Saluga. 2013. “Coal Resource Base in Poland from the Perspective of Using the JORC Code” 23rd World Mining Congress 2013 Proceedings. doi:10.13140/2.1.5119.8724. Somanathan, E., T. Sterner, T. Sugiyama, D. Chimanikire, N.K. Dubash, J. Essandoh-Yeddu, S. Fifita, et al. 2014. “National and Sub-National Policies and Institutions.” In Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, edited by O. Edenhofer, R. Pichs-Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, et al. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press. Speight, James G. 2012. The Chemistry and Technology of Coal. 3rd ed. Boca Raton, USA: CRC Press. Stackelberg, Heinrich von. 1934. Marktform und Gleichgewicht. Wien und Berlin: J. Springer. Stanger, Rohan, Terry Wall, Reinhold Spörl, Manoj Paneru, Simon Grathwohl, Max Weidmann, Günter Scheffknecht, et al. 2015. “Oxyfuel Combustion for CO2 Capture in Power Plants.” International Journal of Greenhouse Gas Control, Special Issue commemorating the 10th year anniversary of the publication of the Intergovernmental Panel on Climate Change Special Report on CO2 Capture and Storage, 40 (September): 55–125. doi:10.1016/j.ijggc.2015.06.010. Statoil. 2016. “Energy Perspectives 2016: Long-Term Macro and Micro Outlook.” Stavanger, Norway: Statoil ASA. http://www.statoil.com/no/NewsAndMedia/News/2016/Downloads/Energy%20Perspectives%2 02016.pdf. Stavins, Robert N. 2003. “Chapter 9 - Experience with Market-Based Environmental Policy Instruments.” In Handbook of Environmental Economics, edited by Karl-Göran Mäler and Jeffrey R. Vincent, 1:355–435. Environmental Degradation and Institutional Responses. Elsevier. http://www.sciencedirect.com/science/article/pii/S1574009903010143. Stechow, Christoph von, Jim Watson, and Barbara Praetorius. 2011. “Policy Incentives for Carbon Capture and Storage Technologies in Europe: A Qualitative Multi-Criteria Analysis.” Global Environmental Change, Special Issue on The Politics and Policy of Carbon Capture and Storage, 21 (2): 346–57. doi:10.1016/j.gloenvcha.2011.01.011. Steckel, Jan Christoph, Ottmar Edenhofer, and Michael Jakob. 2015a. “Drivers for the Renaissance of Coal.” Proceedings of the National Academy of Sciences 112 (29): E3775–81. doi:10.1073/pnas.1422722112. ———. 2015b. “Drivers for the Renaissance of Coal.” Proceedings of the National Academy of Sciences 112 (29): E3775–81. doi:10.1073/pnas.1422722112.
181 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Sterner, Thomas. 2012. “Distributional Effects of Taxing Transport Fuel.” Energy Policy 41 (February): 75–83. doi:10.1016/j.enpol.2010.03.012. Stiglitz, Joseph E. 2015. “Overcoming the Copenhagen Failure with Flexible Commitments.” Economics of Energy & Environmental Policy 4 (2). doi:10.5547/2160-5890.4.2.jsti. SUEK. 2011. “SUEK Annual Report 2011.” Open Joint Stock Company Siberian Coal Energy Company. Sussams, L., and A. Grant. 2015. “The US Coal Crash – Evidence for Structural Change.” Carbon Tracker Initiative. http://www.carbontracker.org/report/the-us-coal-crash/. Tahvonen, Olli. 1995. “International CO₂ Taxation and the Dynamics of Fossil Fuel Markets.” International Tax and Public Finance 2 (2): 261–78. doi:10.1007/BF00877501. Tavoni, Massimo, and Bob van der Zwaan. 2011. “Nuclear Versus Coal plus CCS: A Comparison of Two Competitive Base-Load Climate Control Options.” Environmental Modeling and Assessment 16: 431–440. Tazazanov, Igor. 2012. “Analytical Review Russian Coal Industry.” Ugol Magazine, no. Special Issue: 3–13. Tewalt, Susan J., Harvey E. Belkin, John R. SanFilipo, Matthew D. Merrill, Curtis A. Palmer, Peter D. Warwick, Alexander W. Karlsen, Robert B. Finkelman, and Andy J. Park. 2010. “Chemical Analyses in the World Coal Quality Inventory, Version 1.” 2010–1196. U.S. Geological Survey Open-File Report. U.S. Geological Survey. http://pubs.usgs.gov/of/2010/1196/index.html. The Crown Estate, Carbon Capture & Storage Association, and DECC. 2013. “CCS Cost Reduction Taskforce - Final Report.” London, UK: UK Carbon Capture and Storage Cost Reduction Task Force. The Parliament of Great Britain. 2013. Energy Bill. HL Bill 30. The Stationary Office, London, UK. The State Council of the People’s Republic of China. 2016. “Coal Capacity Guideline Issued.” The State Council of the People’s Republic of China. February 5. http://english.gov.cn/policies/latest_releases/2016/02/05/content_281475284701738.htm. Thurber, Mark C., and Richard K. Morse. 2015a. The Global Coal Market: Supplying the Major Fuel for Emerging Economies. Cambridge University Press. ———. 2015b. The Global Coal Market: Supplying the Major Fuel for Emerging Economies. Cambridge, UK: Cambridge University Press. Tong, Anote. 2015. “Moratorium on New and Extension of Existing Coal Mines.” Trüby, Johannes. 2013. “Strategic Behaviour in International Metallurgical Coal Markets.” Energy Economics 36: 147–57. doi:10.1016/j.eneco.2012.12.006. Trüby, Johannes, and Moritz Paulus. 2012. “Market Structure Scenarios in International Steam Coal Trade.” The Energy Journal 33 (3): 91–123. doi:10.5547/01956574.33.3.4. Tzimas, Evangelos. 2009. “The Cost of Carbon Capture and Storage Demonstration Projects in Europe.” EUR 24125 EN-2009. Petten, The Netherlands: JRC Scientific and Technical Reports, European Commission. Umbach, Frank. 2015. “The Future Role of Coal: International Market Realities vs. Climate Protection?” 6. Strategy Paper. London, UK: The European Centre for Energy and Resource Security (EUCERS) at the Department of War Studies at King’s College London. http://www.kcl.ac.uk/sspp/departments/warstudies/research/groups/eucers/strategy-paper- 6.pdf. UN. 1997. Kyoto Protocol to the United Nations Framework Convention on Climate Change. Doc FCCC/CP/1997/7/Add.1. Vol. 37 I.L.M. 22 (1998). ———. 2013. “United Nations Framework Classification for Fossil Energy and Mineral Reserves and Resources 2009 Incorporating Specifications for Its Application (ECE ENERGY SERIES No. 42.” New York and Geneva: The United Nations Economic Commission for Europe. ———. 2015a. “Transforming Our World: 2030 Agenda for Sustainable Development.” A/RES/70/1. United Nations.
182 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
———. 2015b. Adoption of the Paris Agreement. Proposal by the President. Draft Decision -/CP.21. FCCC/CP/2015/L.9/Rev.1. Vol. Framework Convention on Climate Change. United Nations Secretary-General’s High-Level Panel on Global Sustainability. 2012. “Resilient People, Resilient Planet: A Future Worth Choosing, Overview.” New York, USA: United Nations. uscib.org/docs/GSPReportOverview_A4%20size.pdf. Van de Graaf, Thijs, and Aviel Verbruggen. 2015. “The Oil Endgame: Strategies of Oil Exporters in a Carbon-Constrained World.” Environmental Science & Policy 54: 456–62. doi:10.1016/j.envsci.2015.08.004. VDKI. 2008. “Jahresbericht 2008. Fakten Und Trends 2007/2008.” Hamburg, Germany: Verein der Deutschen Kohleimporteure e.V. ———. 2011. “Jahresbericht 2011. Fakten Und Trends 2010/2011.” Hamburg, Germany: Verein der Deutschen Kohleimporteure e.V. ———. 2012. “Jahresbericht 2012. Fakten Und Trends 2011/2012.” Hamburg, Germany: Verein der Deutschen Kohleimporteure e.V. ———. 2013. “Jahresbericht 2013. Fakten Und Trends 2012/2013.” Hamburg, Germany: Verein der Deutschen Kohleimporteure e.V. VGB PowerTech. 2011. “Geplante Neubauprojekte in Der EU.” Essen, Germany. http://www.vgb.org/neubauprojekte.html. Viebahn, Peter, Daniel Vallentin, and Samuel Höller. 2014. “Prospects of Carbon Capture and Storage (CCS) in India’s Power Sector – An Integrated Assessment.” Applied Energy 117: 62–75. doi:10.1016/j.apenergy.2013.11.054. Warrick, Joby, and Juliet Eilperin. 2016. “Obama Announces Moratorium on New Federal Coal Leases.” The Washington Post. January 15. Weitzman, Martin L. 2015. “Internalizing the Climate Externality: Can a Uniform Price Commitment Help?” Economics of Energy & Environmental Policy 4 (2). doi:10.5547/2160-5890.4.2.mwei. Western Interstate Energy Board. 2015. “Coal Ports.” Accessed October 8. http://westernenergyboard.org/topics/spotlight/coal-ports/. World Bank. 1997. “Five Years after Rio: Innovations in Environmental Policy.” Environmentally & Socially Sustainable Development Work in Progress 18*ESSD. Environmentally Sustainable Development Studies and Monographs Series. Washington, D.C., USA: The World Bank. http://documents.worldbank.org/curated/en/1997/06/694732/five-years-after-rio-innovations- environmental-policy. World Energy Council. 2013. “World Energy Resources: 2013 Survey.” London: World Energy Council. World Summit of the Regions. 2014. “The Paris Declaration - Opportunity for Bottom-Up Action in Favour of the Low Carbon Economy.” Paris, France: Regions of Climate Action. WorleyParsons, and Schlumberger. 2011. “Economic Assessment of Carbon Capture and Storage Technologies: 2011 Update.” The Global CCS Institute. Camberra, Australia. WTO. 1994. Agreement on Subsidies and Countervailing Measures. https://www.wto.org/english/docs_e/legal_e/24-scm_01_e.htm. Xue, Hao, Hanjie Wang, Richard Bridle, Ivetta Gerasimchuk, and Clement Attwood. 2015. “Subsidies to Coal Production in China.” Global Subsidies Initiative. Geneva, Switzerland: International Institute for Sustainable Development (IISD).
ZEP. 2011a. “The Costs of CO2 Capture. Post Demonstration CCS in the EU.” European Technology Platform for Zero Emissions Fossil Fuel Power Plants. Zero Emissions Platform, Brussels, Belgium.
———. 2011b. “The Costs of CO2 Transport. Post Demonstration CCS in the EU.” Brussels, Belgium: European Technology Platform for Zero Emissions Fossil Fuel Power Plants. Zero Emissions Platform. http://www.zeroemissionsplatform.eu/downloads/813.html.
183 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Zhang, Yujiang, Guorui Feng, Min Zhang, Hongrui Ren, Jinwen Bai, Yuxia Guo, Haina Jiang, and Lixun Kang. 2016. “Residual Coal Exploitation and Its Impact on Sustainable Development of the Coal Industry in China.” Energy Policy 96 (September): 534–41. doi:10.1016/j.enpol.2016.06.033.
184 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Appendix A MATHEMATICAL FORMULATIONS AND ADDITIONAL DATA FOR CHAPTER 2, 3, 4
A.1 Model Details: mathematical formulation, node structure, data, and results This Appendix details the optimization problems of all player types as well as their KKTs (Karush Kuhn Tucker conditions). The KKT optimality conditions of each model player and the additional final demand, market clearing and quality equations form a mathematical equilibrium problem in the MCP format. This model is programmed in GAMS and it is solved using the PATH solver (Ferris and Munson, 2000).
A.1.1 Set, parameters, and variables Table A.1, Table A.2, and Table A.3 provide a full nomenclature of all sets, parameters and variables used in the model.
Table A.1: List of sets in the COALMOD-World model. Set name Description Range [2010, 2015, 2020, a model year 2025, 2030, 2035, 2040, 2045, 2050] c consumer see Table A.4 e exporter see Table A.4 f producer see Table A.4
Table A.2: List of parameters in the COALMOD-World model. Parameter name Description Unit E cape initial export capacity of exporter e [Mt/a] P cap f initial production capacity of producer [Mt/a] TC cap fc initial transport capacity from producer to consumer c [Mt/a] TE cap fe initial transport capacity from producer to exporter e [Mt/a] China_ lic Chinese export license volume [Mt/a] aECHN E investment cost for export capacity expansion for exporter [USD/t] Cinvae e P investment cost for producer capacity expansion for Cinv [USD/t] af producer
185 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
TC investment cost for transport capacity expansion from Cinv [USD/t] afc producer f to consumer c investment cost for transport capacity expansion from CinvTE [USD/t] afe producer to exporter e
feeae port handling fee for exporter e [USD/t] E [Mt/a per 5 year maximum export capacity expansion of exporter e invae period] P [Mt/a per 5 year maximum production capacity expansion of producer invaf period]
e energy content of coal shipped by exporter e [t/GJ]
f energy content of coal produced by producer [t/GJ] mc__ int start af slop of marginal cost curve for producer USD/t
mc__ int varaf intercept variation factor (mine mortality rate) USD/t mc__ slop start af starting value of marginal cost intercept for producer USD/t
plength period length 5 years
re discount factor applied by exporter e [%]
rf discount factor applied by producer [%]
res f resource endowment of producer [Mt]
searateaec freight rate for transport from exporter e to consumer [USD/t] C transafc transportation cost from producer to consumer [USD/t] E transafe transportation cost from producer to exporter e [USD/t]
Table A.3: List of variables in the COALMOD-World model. Variable name Description Unit
TC shadow price of transport capacity constraint from cap [USD/t] aafc producer to consumer c
TE shadow price of transport capacity constraint from cap [USD/t] aafe producer to exporter e
P shadow price for maximal production capacity expansion inv [USD/t] aaf constraint for producer shadow price of production capacity constraint for P [USD/t] af producer res f shadow price of resource constraint [USD/t] E invae investment in export capacity by exporter e in period a [Mt/a]
P investment in production capacity by producer in inv [Mt/a] af period
TC investment in transport capacity from producer to inv [Mt/a] afc consumer c in period
TE investment in transport capacity from producer to inv [Mt/a] afe exporter e in period C pac price paid by consumer to exporter or producer [USD/GJ] E pae price paid by exporter to producer [USD/GJ]
186 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
E a export-based tax [USD/GJ] P a production-based tax [USD/GJ]
xafc Sales from producer f to consumer c [GJ/a]
yafe Sales from producer to exporter e [GJ/a]
zaec Sales from exporter e to consumer c [GJ/a]
A.1.2 Producer formulation Optimization Problem
The producer maximizes its profit under reserve constraints and technical restrictions on its production and land transport capacity in every year.
In the second line of the producers' objective function (1) we can see the summation of the yearly net revenues in the squared brackets over all model years with the associated discount rate rf . The following two terms after the brackets are the revenues from sales to local demand nodes and to exporters. The third line of (1) shows the production cost function in an undefined form. The fourth line of (1) represents the transport costs to local demand and exporters. Line five of (1) calculates the total investment costs in production capacity and line six does the same for the investments in transport capacities to local demand and exporters.
In case of a producer tax which is introduced in4.4.2, tax payments also needs to be considered by the producer which is done in line six of (1).
P P TC TE maxf (x afc ; y afe ; inv ; inv ; inv ) x;;;; y invP inv TC inv TE af afc afc afc afe af afc afc a 1 CE [ p xafc p ae y afe a A cac e 1 rf P xy, Caf afc afe CE trans f x afc trans afe f y afe (A.1) ceafc PP Cinvaf inv af TC TC TE TE Cinvafcinv afc Ci n v afe inv afe ce xy P afc afe a ] c e s.t.
capPPP inv x y 0 (A.2) f af f afc f afe af a a c e
P P invP af (A.3) inv invaf 0 af
187 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
resf[( f x afc f y afe a startyear c e (A.4) plength res f xy( a 1) fc f ( a 1) fe)]·0 f ce 2
TC TC capTC capfc inv afc f x afc 0 afc (A.5) aa
TE TE capTE capfe inv afe f y afe 0 afe (A.6) aa
P TC TE xafc0; y afe 0; inv af 0; inv afc 0; inv afe 0 (A.7)
Equation (A.2) represents the production capacity constraint for one year which depends on the capacity in the starting year and investments in subsequent periods prior to the model year. Equation (A.3) is a restriction on the maximum investments in production capacity that can be build up during the next five years (i.e. until the next model year). Equation (A.4) is the reserve constraint of the producer over the total model running time and includes reserve utilization from the production of the years between the model years. On the domestic transport market we have (A.5) and (A.6) which are the transport capacity constraints for each model year for transport routes to local demand nodes and exporters, respectively. The symbols in parentheses are the dual variables associated with the constraints and (A.7) are the non-negativity constraints of the decision variables. The endogenous cost mechanism that enters the model with the following equation:
mc__ intaf mc int( a 1) f
mc___ slp mc int var x y , (A.8) (a 1) f f f ( a 1) fc ( a 1) fe ce
mc_ intaf (free)
Equation (A.8) states that the intercept in year a is equal to the previous period's intercept plus the previous period's slope multiplied by the production in that year and the factor mc_ int _ varf [0,1].
The mine mortality factor mc__ int varf determines how fast the cheapest mines are mined out.
Optimality Conditions (KKTs)
The producer’s profit maximization problem has the following Karush-Kuhn-Tucker conditions (KKTs) of optimality that are obtained after deriving the Lagrangian function for each producer with respect to its decision variables and dual variables of constraints. The red components of the equations only apply in case of a producer tax, which is for instance introduced in4.4.2.
a 1 C P 0 [] pCC af trans 1ryac afc f f afe (A.9)
plength TC P res cap P x 0 af f2 f f afcf a af c
188 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
a 1 C P 0 [] pEE af trans 1ryae afe f f afe (A.10)
plength TE P res cap P y 0 af f2 f f afef a af e
a 1 TC capTC inv TC TC 00Cinvafc afc afc inv afc (A.11) aa 1 rf
a 1 TE capTE inv TE TE 00Cinvafe afe afe inv afe (A.12) aa 1 rf
00capPPP inv x y (A.13) f af f afc f afe af a a c e
P P invP 00invaf invaf af (A.14)
0 resf [( f x afc f y afe a startyear c e (A.15) plength res f xy( a 1) fc f ( a 1) fe)]·0 f ce 2
TC TC capTC 00capfc inv afc f x afc afc (A.16) aa
TE TE capTE 00capfe inv afe f y afe afe (A.17) aa
0mc _ intaf mc _ int( a 1) f
mc_ slp mc__ int var x y , (A.18) (a 1) f f f ( a 1) fc ( a 1) fe ce
mc_ intaf (free)
A.1.3 Exporter formulation Optimization Problem
The exporter's decision is to choose the optimal quantity zaec to sell to each importing country c in
E each year a and also to invest in export capacity invae .
In case of an export tax which is introduced in Chapter 4, tax payments also needs to be considered by the exporter which is done in line four of (A.19).
189 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
a EE1 maxe(;) z aec inv ae E aA zaec; inv ae 1 re CE [pac z aec p ae z aec fee ae e z aec c c c
EE searateaece z aec Cinv ae inv ae c E a z aec' ] (A.19) s.t.
EE E cape invae e zaec 0 ae (A.20) a a c
E E invE ae (A.21) inv invae 0 ae
E zaec0; inv ae 0 (A.22)
Constraint (A.20) represents the maximum export capacity in each model year which depends on the capacity in the starting year and investments in subsequent periods prior to model year a . Equation (A.21) expresses the maximum investments in export capacity for one model period. The symbols in parentheses are the dual variables associated with the constraints.
Equation (A.23) is the clearing of sales of the producer to the exporter. Its dual variable gives the price
E at which the exporter receives the coal from the producer, pae .
E 0,yafe z aec p ae (free) (A.23) c
Modeling China's export restriction requires an additional equation. Chinese coal exports are restricted by politically determined export licenses. Thus we put a constraint on all consumption nodes with a non-Chinese import port (i.e., countries NoChina c ) using equation (A.24). China_ lic aECHN represents the level of Chinese export licenses for a given year in million tons.
China_0 licaE z aec ae aE (A.24) CHNNoChina() c CHN
Optimality Conditions (KKTs)
a 1 CE E 0· pac p ae fee ae e searate aec e a 1 re (A.25) E ·0 z aee ec aECHN e a ec
190 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
a 1 E E invE E 0 ·Cinvae ae ae inv ae 0 (A.26) aa 1 re
EE E 00cape invae e z aec ae (A.27) aca
E E invE 00invae invae ae (A.28)
0China _ licaE z aec e aE 0 (A.29) CHNNoChina() c CHN
A.1.4 Final demand formulation Market Clearing
The following market clearing condition determines the price given the demand function
C pac x afc, z aec at the demand node c .
CCC pac p ac x afc, z aec 0 , p ac (free) (A.30) fe
A.2 Model extension: Policy maker as Stackelberg-leader at the upper level
A.2.1 Export tax To extend the model framework one can for instance add an upper level, where a policy maker g can
E levy a tax a on steam coal exports in periods aA in order to maximize the NPV of tax revenues.
E While the policy maker can decide on the initial tax rate 0 starting in period aa , the path is
110 predetermined by an annual growth rate of r :
EE aa a 0 (1 r ) (A.31)
Accordingly, the policy maker's optimization problem is given by
a 1 E max aEXP aec' (A.32) E 0 aec' 1 rg
where EXPaec subsumes total exports of all exporters e being located in decision maker ’s territory, and directed towards consumption nodes c , which are outside ’s territory. It is the impact
110 By varying this growth rate, the analysis is not restricted to a constant tax rate but rather allows for an endogenous starting level that gradually increases over time. We provide sensitivity analyses of the revenue- maximizing tax rate regarding the growth rate of the tax in Appendix A.5. Note that the alternative to implement the tax rate as a free variable for each period could create time inconsistencies.
191 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
of the tax on total exports that can be anticipated by the policy maker. Periodic revenues are discounted at rate rg . Note that we model the export tax based on the energy content of exported volumes; it is hence proportional to a carbon tax.
A.2.2 Production tax
P The tax path in Equation. (A.31) now holds for production a :
PP aa a 0 (1 r ) (A.33)
Accordingly, the optimization problem given by Eq. (A.32) is adjusted to
a 1 P max aPROD af , (A.34) P 0 af 1 rg
where PRODaf subsumes the production of all producers f being in the territory of policy maker g are affected by the production tax. The producer’s maximization problem from Equation (A.1) is adjusted by incorporating the tax rate, accordingly.
192 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.4: Nodes of COALMOD-World.
Country Producers Regions Exporters Port Consumers Regions Port Australia P_AUS_QLD Queensland E_AUS_QLD Gladstone C_AUS No P_AUS_NSW New South Wales E_AUS_NSW Botany Bay Brazil C_BRA Fortaleza Canada P_CAN C_CAN Ontario No Chile C_CHL Mejillones China P_CHN_SIS Gansu, Inner Mongolia, Hebei, Ningxia, Shaanxi, E_CHN Qinhuangdao C_CHN_Northeas Heilongjiang, Jilin, Liaoning No Shaaxi, Shanxi t P_CHN_Northeast Heilongjiang, Jilin, Liaoning, Tianjin C_CHN_Main Beijing, Tianjin, Hebei, Henan, Shandong, Tianjin No P_CHN_HSA Anhui, Bejing, Fujian, Henan, Jiangsu, Jiangxi, C_CHN_Eastern Anhui, Jiangsu, Hubei, Shanghai, Zhejiang Shanghai/Ningbo Shandong, Zhejiang P_CHN_YG Chongqing, Guangxi, Guizhou, Guizhou, Hubei, C_CHN_South Chongqing, Fujian, Guangdong, Guangxi, Guizhou, Hong Guangzhou Hunan, Sichuan, Yunnan Kong, Hunan,Jiangxi, Sichuan C_CHN_SIS Shanxi, Shaaxi, Inner Mongolia No Colombia P_COL E_COL Cartagena Denmark C_DNK Aalborg Finland C_FIN Kotka Germany C_DEU Rotterdam + land transport India P_IND_North Jharkhand, Madhya Pradesh, Chhattisgarh, West C_IND_East Bihar, Chhatisgarh, Jharkhand, Orissa, West Bengal No Bengal P_IND_Orissa Orissa C_IND_North Delhi, Punjab, Rajasthan, Uttar Pradesh No P_IND_West Maharashtra C_IND_West Gujarat, Madhya Pradesh, Maharashtra Kandla P_IND_South Andhra Pradesh C_IND_South Andhra Pradesh, Karnataka, Tamil Nadu Chennai Indonesia P_IDN E_IDN Surabaya C_IDN No Israel C_ISR Ashdod Italy C_ITA Taranto Japan C_JPN Yokohama Kazakhstan P_KAZ Ekibastuz E_RUS_West Ust‐Luga C_KAZ No Korea C_KOR Ulsan Malaysia C_MYS Port Klang Mexico C_MEX Manzanillo Mongolia P_MNG Morocco C_MAR Mohammedia Mozambique P_MOZ E_MOZ Maputo Netherlands/ C_NFB Netherlands, France, Belgium Rotterdam France/Belgium Philippines C_PHL Manila Poland P_POL E_POL Gdansk C_POL No Portugal C_PRT Sines Russia P_RUS Kemerovo/Kuznets E_RUS_East Vladivostok C_RUS_Central No E_Black_Sea_RUS Odessa C_RUS_Siberia No South Africa P_ZAF E_ZAF Richards Bay C_ZAF No Spain C_ESP Gijon Taiwan C_TWN Kaohsiung Thailand C_THA Bangkok Turkey C_TUR Mersin + Black sea port UK C_GBR Immingham Ukraine P_UKR Ukrainian/Russian Donetsk E_Black_Sea_UKR Odessa C_UKR No USA P_USA_PRB Powder River Basin E_USA_SC Houston, TX C_USA_W AK, AZ, CA, CO, ID, MT, NV, NM, OR, UT, WA,WY No P_USA_Rocky Rocky Mountains E_USA_SE Norfolk, VA C_USA_NC IL, IN, IA, KS, MI, MN, NE, MO, ND, OH, SD, WI No P_USA_ILL Illinois Basin E_USA_W Portland, OR C_USA_SC AR, LA, OK, TX No P_USA_APP Appalachian C_USA_SE AL, DE, DC, FL, GA, KY, MD, MS, NC, SC, TN, VA, WV Mobile C_USA_NE CT, ME, MA, NH, NJ, NY, PA, RI, VT No Venezuela P_VEN E_VEN Puerto la Cruz Vietnam P_VNM
193 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.5: World Energy Outlook demand projections for coal for power generation in the scenarios (Mtoe). New Policies Scenario 2°C scenario Regions 2013 2020 2030 2040 2020 2030 2040 Data origin WORLD 3929 4033 4219 4414 3752 2889 2495 TPED coal OECD Americas 428 362 293 258 294 144 161 Coal in Electricity USA 397 333 272 242 269 136 156 Coal in Electricity OECD Europe 230 194 120 75 171 57 41 Coal in Electricity European Union 218 179 101 58 159 55 41 Coal in Electricity OECD Asia Oceania 165 149 130 107 139 62 26 Coal in Electricity Japan 70 62 58 50 57 24 5 Coal in Electricity E. Europe / Eurasia 140 130 122 122 118 61 41 Coal in Electricity Russia 70 67 69 67 59 31 23 Coal in Electricity Non OECD Asia 1368 1550 1798 2030 1416 991 725 Coal in Electricity China 2053 2060 2078 1978 1966 1542 1247 TPED coal India 341 476 690 934 442 459 453 TPED coal Africa 64 71 81 97 56 52 42 Coal in Electricity South Africa 95 94 90 85 91 71 56 TPED coal Latin America 24 29 38 46 28 25 26 TPED coal Brazil 16 20 23 26 19 16 15 TPED coal Source: IEA (2015a).
Table A.6: Various input parameters for COALMOD-World production nodes. Prod. Invest- Max. capacity Prod. cost Prod. cost Mine capacity ment expansion per 5 function intercept function mortality 2010 Reserves costs year 2010 slope rate (Mtpa) (Mt) ($/tpa) (Mtpa) ($/t) ($/t*t) (%) US PRB 525 112,555 40 100 10 0.019 8 US Rocky 79 20,704 60 26 30 0.0633 1 US ILL 115 82,887 50 34 40 0.1128 3 US APP 336 54,572 70 64 40 0.119 10 Colombia 75 6,229 50 30 39 0.1333 15 Venezuela 10 479 50 10 41 1.346 6 Poland 71 13,997 80 5 41 0.831 2 Ukraine 45 16,271 70 10 31 0.6889 7 Kazakhstan 100 28,145 40 15 15 0.147 20 Russia 190 49,078 50 51 15 0.1737 4 South Africa 267 48,740 50 40 20 0.1124 2 IND North 281 35,663 40 60 9 0.1709 1 IND Orissa 123 14,416 40 30 11 0.1302 2 IND West 53 7,134 40 10 22 0.0759 5 IND South 58 6,755 40 10 23 0.344 5 Vietnam 62 150 60 10 12 0.327 60 Indonesia 340 13,000 35 90 23 0.0794 3 CHN SIS 1573 213,400 60 300 21 0.0267 0.8 CHN Northeast 121 15,900 60 50 26 0.2 1 CHN HSA 564 4,700 60 70 30 0.0514 1 CHN YG 450 36,800 60 30 25 0.1111 14 AUS QLD 85 24,764 40 20 25 0.5882 5 AUS NSW 119 13,829 50 20 30 0.3782 8 Mongolia 17 1,170 60 20 30 0.2 10 Mozambique 5 212 80 15 45 0.05 10
194 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.7: Results of COALMOD-World: consumption, domestic supply, and imports by consuming country and scenario in 2010, 2020, 2030, and 2040.
Stagnation scenario [Mt] 2°C scenario [Mt]
2010 2020 2030 2040 2020 2030 2040
Dest Imp Dom Imp Dom Imp Dom Imp Dom Imp Dom Imp Dom Imp Dom AUS 0 63 0 57 0 50 0 41 0 53 0 24 0 10 BRA 5 0 6 0 7 0 9 0 5 0 5 0 5 0 CAN 9 0 8 0 6 0 6 0 6 0 3 0 4 0 CHL 8 0 8 0 11 0 13 0 8 0 7 0 8 0 CHN 95 2404 229 2285 278 2229 168 2200 218 2182 313 1567 286 1239 DEU 31 0 26 0 14 0 8 0 23 0 8 0 6 0 DNK 4 0 4 0 2 0 1 0 3 0 1 0 1 0 ESP 10 0 8 0 4 0 3 0 7 0 2 0 2 0 FIN 5 0 3 0 2 0 1 0 3 0 1 0 1 0 GBR 20 0 15 0 9 0 5 0 14 0 5 0 4 0 IDN 0 58 0 65 0 76 0 86 0 60 0 42 0 31 IND 63 488 165 587 326 740 469 945 135 567 157 569 171 541 ISR 9 0 8 0 4 0 2 0 7 0 2 0 2 0 ITA 14 0 12 0 7 0 4 0 11 0 4 0 3 0 JPN 123 0 104 0 97 0 85 0 96 0 40 0 8 0 KAZ 0 61 0 56 0 53 0 53 0 51 0 27 0 18 KOR 83 0 70 0 61 0 50 0 64 0 29 0 12 0 MAR 4 0 5 0 5 0 6 0 4 0 3 0 3 0 MEX 9 0 9 0 12 0 15 0 9 0 8 0 8 0 MYS 21 0 24 0 27 0 31 0 21 0 15 0 11 0 NFB 22 0 18 0 10 0 6 0 16 0 6 0 4 0 PHL 11 0 12 0 12 0 14 0 11 0 8 0 5 0 POL 0 71 0 58 0 33 0 19 0 52 0 18 0 13 PRT 4 0 4 0 2 0 1 0 3 0 1 0 1 0 RUS 30 69 36 59 33 64 29 65 24 59 19 25 14 19 THA 17 0 16 0 18 0 21 0 15 0 10 0 7 0 TUR 16 0 15 0 14 0 14 0 14 0 7 0 5 0 TWN 62 0 62 0 68 0 79 0 60 0 40 0 28 0 UKR 19 19 30 5 30 3 30 3 26 5 16 0 11 0 USA 0 886 0 762 0 646 0 586 0 624 0 332 0 378 ZAF 0 187 0 185 0 177 0 167 0 179 0 140 0 110
Table A.8: Results of COALMOD-World: domestic supply and exports by producing country and scenario in 2010, 2020, 2030, and 2040.
Stagnation scenario [Mt] 2°C scenario [Mt]
2010 2020 2030 2040 2020 2030 2040
Prod Exp Dom Exp Dom Exp Dom Exp Dom Exp Dom Exp Dom Exp Dom AUS 93 63 105 57 115 50 119 41 94 53 78 24 58 10 CHN 20 2404 8 2285 0 2229 0 2200 51 2182 0 1567 0 1239 COL 60 0 112 0 136 0 139 0 70 0 60 0 43 0 IDN 270 58 284 65 285 76 237 86 272 60 270 42 238 31 IND 0 488 0 587 0 740 0 945 0 567 0 569 0 541 KAZ 30 61 36 56 33 53 29 53 24 51 19 27 14 18 MNG 17 0 43 0 42 0 16 0 35 0 29 0 23 0 MOZ 4 0 10 0 9 0 0 0 9 0 7 0 0 0 POL 0 71 0 58 8 33 23 19 0 52 1 18 0 13 RUS 71 69 102 59 106 64 108 65 87 59 77 25 67 19 UKR 15 19 34 5 34 3 34 3 29 5 29 0 17 0 USA 19 886 48 762 156 646 207 586 21 624 8 332 9 378 VEN 5 0 10 0 14 0 14 0 9 0 8 0 3 0 VNM 18 0 0 0 0 0 0 0 0 0 0 0 0 0 ZAF 72 187 106 185 126 177 146 167 102 179 126 140 136 110 Sum 694 898 1064 1072 803 712 608
195 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.9: Trade flows in COALMOD-World (in Mtpa). 2010 2020 2030 2040 From To NPS 450 NPS 450 NPS 450
USA Canada 9 8 6 6 3 6 4 Appalachia South Europe 65 28 43 0 0 0 0 America Russia Kazakhstan 30 36 24 33 19 29 14 (Central) Russia & Ukraine 18 30 26 30 16 30 11 Ukraine Russia Europe 23 18 10 28 20 4 18 Russia Turkey 26 23 21 19 10 16 7 Russia OECD Asia 20 36 35 36 3 49 0 (via Far East) Europe & South Africa West 22 4 3 0 0 0 0 Mediterranean South Africa India 50 102 99 126 126 146 136 Vietnam China 18 0 0 0 0 0 0 Indonesia India 9 53 27 139 23 206 34 Asia Indonesia 244 231 245 147 247 31 204 (except India) China S. Korea 20 8 51 0 0 0 0 India & Australia 0 16 15 18 10 21 7 Thailand Asia (except Australia India & 93 89 79 97 68 98 51 Thailand) Mongolia China 17 43 35 42 29 16 23 Mozambique India 4 10 9 9 7 0 0 South Asia 0 94 24 150 67 152 43 America USA West Asia 0 0 0 125 0 199 0 USA Europe 5 17 4 2 0 0 0 Appalachia Poland Europe 0 0 0 8 1 23 0 Russia China 0 0 0 0 32 0 30 Europe & Black Sea West 0 28 25 19 11 10 5 Mediterranean Chile, Brazil, Indonesia 17 0 0 0 0 0 0 Mexico Chile, Brazil, Russia 0 0 0 8 15 34 13 Mexico Chile, Brazil, USA West 5 23 10 22 5 3 5 Mexico South Chile, Brazil, 0 0 13 0 1 0 3 America Mexico
196 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.10: Comparison of key statistics across scenarios for at the tax-revenue maximizing level, and at 10 USD/t as the initial tax rate.
NPV of Average Revenue- tax reduction in Average Reduction Average maximising revenue production of reduction in in global Reduction emissions Abatement Scenario initial tax 2015- tax setting global production in Average reduction Leakage Tax level 2035 [bn countries production 2015-2035 seaborne price 2015-2035 rate; CO2- Revenue [USD/tCO2] USD] [Mt/a]* [Mt/a]* [%]* trade [%] change [%] [GtCO2/a]* based [%] [USD/tCO2] Optimal tax level Tax AUS
Export tax 6.7 16 52 12 0.20 3.2 0.8 0.04 73.3 22 Production tax** 8.8(39.4) 32(44) 77(155) 26(69) 0.44(1.17) 4.0(6.2) 1.5(3.5) 0.07(0.19) 63.5(52.4) 22(11)
Tax Coalition
Export tax 10.1 125 311 84 1.42 20.3 3.2 0.19 74.7 32 Export tax + USA 12.1 151 543 193 3.26 35.5 6.7 0.39 69.8 19 Production tax 12.2 266 426 180 3.04 22.4 6.4 0.42 59.4 31
Grand tax coalitions
All exporters + Export tax 42.6 344 567 513 8.67 74.9 15.8 1.22 35.9 14 All exporters + 26.0 858 1428 1098 18.55 55.1 34.4 2.63 23.9 16 Production Tax All producers + Production 40.5 3634 3095 3095 52.30 50.1 111.6 7.61 0 24 tax
Climate Scenario*** n.a n.a 27.3 30.8% -17% 4.0 At 10 USD/tCO2 initial tax rate Tax AUS Export tax 11 34 91 24 0.41 5 1.6 0.072 70.1 7 Production tax 32 0(27) 99 33 0.56 4.9 2.0 0.096 63.2 17
Tax Coalition Export tax 125 0 308 83 1.40 20.1 3.2 0.192 74.7 33 Export tax + USA 150 1 470 163 2.75 29.5 5.6 0.323 71.0 23 Production tax 253 5 340 142 2.40 17.4 5.0 0.336 59.7 38
Grand tax coalitions
All exporters + Export tax 211 39 200 164 2.77 40.8 6.2 0.347 66.7 30 All exporters + Production 569 34 675 407 6.88 29.1 13.5 0.957 41.9 30 Tax All producers + Production 1635 55 777 777 13.13 9.8 26.5 1.904 0 43 tax
197 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.3 MPEC solution algorithm The Disjunctive Constraints formulation is an established method to solve MPECs, where the complementarity conditions of the lower level are reformulated into a MIP. However, this technique has two drawbacks well discussed in the literature: for large models the method is computationally expensive (cf. Luo et al., 1996), while at the same time it requires the definition of upper bounds for all endogenous variables. Results are highly sensitive to these bounds and a ”bad choice” can generate misleading results and infeasibilities (cf. Gabriel and Leuthold, 2010).111 We develop an algorithm that combines NLPEC and Disjunctive Constraints solution techniques to overcome the drawbacks of the individual methods. Figure A.1 depicts how, the two alternative formulations are combined in our approach.
Figure A.1: Illustration of MPEC solution strategy.
A.3.1 First step: GAMS NLPEC solver In a first step, which is similar to the approach described in Chapter 4.4.2, we solve the model using the NLPEC solver. In order to obtain different local optima as candidates for the global optimum, we
e vary the upper and lower bound of decision variable 0 . Furthermore, extreme values for all variables are reported and stored for further use in the second step. To this end the model is defined by the upper-level objective function (A.32) together with the lower level being formulated by means of KKT conditions, as denoted in Equations (A.9)-(A.18), (A.23)., and (A.25)-(A.30).
111 Another method proposed by Siddiqui and Gabriel (2013) uses Schur’s decomposition and variables of specially ordered sets (SOS-Type 1 variables) to avoid exogenously choosing suitable (upper) bounds.
198 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.3.2 Second step: disjunctive constraint reformulation In a second step, we test the candidate NLPEC solution, which has the highest upper-level objective value, for optimality by formulating a linear MIP. For this purpose the upper and lower level needs to be linearized, and the complementarity conditions, i.e. the duality between equation and respective variable, need to be replaced by disjunctive constraints. Both a binary variable bin , and a sufficiently large positive constant have to be defined for each complementarity condition.
For instance, focusing on the exporter’s optimization problem we can write complementarity condition Equation (A.25) by means of disjunctive constraints as follows:
a 1 CEE 0 pac p ae fee e e sea aec e a 1 re (A.35) E bin z K z e aeec aECHN e aec aec
zz 0zaec (1 bin aec ) K aec . (A.36)
Furthermore, the objective function at the upper level, more specifically the bilinear term given by the product of tax rate and exported quantities, has to be linearized. We follow Gabriel and Leuthold
e (2010) in discretising the decision variable to 0,d , where index d denotes predetermined discrete options for the export tax. Each potential tax rate is related to a binary variable bind and a sufficiently large positive constant Kd . In essence, the algorithm determines the highest objective value by choosing one of the given tax rates, i.e. by setting the corresponding binary variable to unity. The linearized upper level expressed by disjunctive constraints reads as follows
obj max revd (A.37) d
aa 1 rev E (1 r ) a a EXP (A.38) d0, d aec aec 1 rg
bind 1 (A.39) d
revd K d bin d rev d K d bin d . (A.40)
While Equation (A.37) maximises the sum of potential tax revenues revd over all possible discrete choices, Equation (A.38) links these revenues to the different tax rate options. By Equation (A.39) it is guaranteed that only one tax options is chosen, while (A.40) only allows the corresponding tax revenue to this tax rate to be positive. This ensures that Equation (A.37) ultimately chooses the highest possible tax revenue.
199 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
In order to test the candidate solution provided by NLPEC, we define the set of exogenous export tax rates in a close range around it. We hence test for optimality in a constrained set of choices.
A.4 Global trade flows in 2030 by case
Figure A.2: Global trade flows in the Base Case in 2030 (in Mt).
Figure A.3: Global trade flows in AUS – Export tax in 2030 (in Mt).
Figure A.4: Global trade flows in Coalition – Export tax in 2030 (in Mt).
200 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.5 Sensitivity analysis: discount rate and tax growth rate The revenue-maximizing initial tax rate is found to be robust to changes in the discount rate within the range of rg [0,0.10]. For instance, a discount rate of 0% changes the revenue-maximizing initial tax rate to 6.72 USD/tCO2 compared to the 6.73 USD/tCO2 with the default assumption of a discount rate at 5%. Moreover, we analyze different predetermined annual growth rates of the export tax, i.e. 0% (constant), 2.5% (default; slow increase), 5% and 10% (fast increasing tax rate) as well as -2.5% (decreasing path). Intuitively, the lower the slope the higher the initial tax value. While the initial revenue-maximizing tax rate decreases monotonically with the growth rate parameter, the function of the NPV of tax revenues is bell-shaped (see Figure A.5). Among the parameters tested, the default setting leads to a slightly lower NPV of tax revenues than a constant tax rate over time.
Figure A.5: AUS – Export tax: Optimal initial tax rates, in USD/tCO2 and the NPV of tax revenues, in bn USD (right axis), as a function of the growth rate of the export tax, in percentage per annum.
A.6 Coalition joint by the USA To check the sensitivity of our results with respect to the members of the coalition, we examine the case with the USA joining the coalition. We find the revenue-maximizing initial tax level at 12.1
USD/tCO2 compared to 10.1 USD/tCO2 in the coalition of the 4 major exporters only.
Total coalition revenues rise to 151 bn USD. While the effect of including the USA into the coalition is quite pronounced - leading to an increase of 2 USD/tCO2 in the revenue maximising tax level, and about 25 bn USD of additional revenues - the US’ share of revenues is rather small compared to the other members (seeFigure A.6). The high tax puts the world coal prices under pressure and leads to an average increase of 6.7%.
Competing exporters and domestic producers are increasingly unable to compensate for the restrained exports which lead to a 3% decrease in global consumption - and thus CO2 emissions -
201 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
over the model horizon. We find that this policy may thus lead to a reduction in total CO2 emissions by more than 7.8 Gt over the next 15 years - twice as much as for a coalition without the USA.
Figure A.6: Coalition + USA – Export tax: NPV of tax revenues of coalition members, in bn. USD; and change in global consumption, export for all coalition members and total exports, in percentages (right axis), as a function of the initial value of the export tax, in USD/tCO2.
202 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.7 Country-by-country assessment of coal production subsidies Australia
Data on current coal production subsidies in Australia is available from different sources which provide very different estimates: while OECD data suggests an average annual subsidy of 0.1 bn USD (2007- 2014) (OECD 2015a), Fulton, Buckley et al. (2015) assume an average direct tax deductions potential that was available from 2005 to 2011 (0.3 bn USD annually), and fuel tax credit scheme available from 2012 to 2013 (0.6 bn USD annually), will both also be available to producers in the future. Additionally, they note that current practice of allowing mining companies to provide less costly financial products as a substitute for rehabilitation bonds, constitute a subsidy to coal mining, which they estimate at 1.5 USD/t or 0.7 bn USD112, annually. Makhijani and Doukas (2015) report national subsidies to coal production of 0.3 bn USD, almost exclusively from direct spending. The 76% of the subsidies are directed towards remediation, while the rest splits between transportation (20%), R&D (3%) and exploration (1%). Some of the subsidies apply on the regional level for production in New South Wales (46%) and Queensland (20%), the remainder applies to all production sites. For the purpose of the analysis in Chapter 3, I employ the conservative values estimated by Makhijani and Doukas (2015), but add the rehabilitation subsidy noted by Fulton et al. (2015), as this constitutes a major subsidy otherwise not covered. The resulting coal production subsidy level is 2.5 USD/t for New South Wales, 2.1 USD/t for Queensland, and 1.8 USD/t for all other regions113.
China
The level of subsidies on coal production in China in 2013 is estimated at 5.8 bn USD excluding 0.6- 5.8 bn USD of support granted through tax credits, which translate to 1.5 to 3 USD/t of coal produced (Xue et al. 2015). These figures also include financial assistance for state-owned enterprises (SOEs) which account for 92% of coal production in 2013 (Xue et al. 2015). Support is granted as part of an industry consolidation and infrastructure improvement plan. While, according to Xue et al. (2015), this subsidy totaled 1.3 bn USD in 2013, ODI (2015a) reports an average support of 6.2 bn USD (2013- 2014). For reasons of consistency, I use the figure from Xue et al. (2015). To arrive at figures to be included in this analysis, I further subtract subsidies for coal-bed methane, which is not covered in the model setting, support for R&D, and oversea investments. Furthermore, I employ the conservative estimate on support through tax credits, which gives a total subsidy level of 4.4 bn USD (originating from state level (71%) and regional level (29%) support). The remainder includes direct payments and investments (54%), provision of services below market value (39%), and subsidies in the form of foregone profits (6%). The resulting calculated subsidy level is 1.4 USD/t for producers from Shanxi, Shaanxi and Inter Mongolia, and 0.9 USD/t for producers from all other regions114.
112 According to IEA (2015c), Australia produced 458 Mt of coal in 2013. 113 According to Queensland Government (2016) coal production in Queensland totaled on average 226 Mt (2012-2015); according to the NSW Department of Industry (2015) coal production in New South Wales totaled 196 Mt in 2013-2014. 114 According to Denjean et al. (2015), in 2013, total coal production in China was 3.7 Gt, with Shanxi, Shaanxi, and Inner Mongolia accounting for 0.96 Gt, 0.493 Gt, and 0.994 Gt, respectively. 203 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A sum of1 bn USD of the subsidies is given as compensation for those coal mines that are shut down due to the coal phase-out plan (Xue et al. 2015), other forms of practiced subsidies require local content or give preferential treatment to state-owned enterprises. According to the definition used in section 3.3.1 of Chapter 3 these payments clearly constitute a subsidy. This example highlights the fact that the definition of a subsidy is non-indifferent on whether it is used to remove market failures or not. However, a detailed evaluation of the efficiency of each individual subsidy is beyond the scope of this study.
India
Data on coal production subsidies in India is only available from Garg and Bossong (2015). They report total average governmental support of 0.8 bn USD for the period 2013-2014 (ODI 2015b). To a small extent, it takes the form of tax breaks and direct funding for exploration, extraction and equipment, but over 90% originates from investment by SOE Coal India Limited (CIL), which account for around 70% of total coal production in India. Taking into account the state’s share and the market share of CIL support translates in a subsidy of 0.9 USD/t of coal produced115.
Indonesia
Lontho and Beaton (2015) undertake a comprehensive effort in compiling fossil fuel subsidies in Indonesia, but find little data available for the coal sector. OCI (2015) report annual government support to coal mining of 0.9 bn USD in 2013, mainly originating from a difference in royalty taxes between small and big mines, and from untaxed production, accounting for 12-15% of annual production (50-90Mt in 2014, Sanzillo 2015). For the royalty tax reforms are announced, but have not been included in any regulation, so far (PwC 2015, 37), and has been impeded by local resistance (Gatot and Sjahrir 2015; Kannan, Das, and Corazon Aureus 2015). Untaxed production is currently targeted by a new policy requiring producers to obtain “clean and clear” certificates (PwC 2015, 10).
For coal produced under the Domestic Market Obligation (DMO) consumers pay a regulated price which is benchmarked against a basket of market-based prices (PwC 2015, 10), including an international reference price. Therefore the case for subsidization via price discrimination cannot be clearly made. Assuming that the two policies equalizing the royalty taxes and preventing unmonitored production remain in their current state and do not become effective, Indonesia exhibits a coal production subsidy of 1.8 USD/t116.
Poland
As Poland is not a G20 country ODI does not provide a country study and thus no fossil fuel subsidy data base. Therefore, the only source that estimates subsidies to coal production in Poland is OECD (2015a). For 2013 and 2014 the data base reports subsidies of 0.76 bn USD which translate to 5.4 USD/t of coal produced using production figures from IEA (2015c). 98% of the subsidy originates from
115 According to IEA (2015c), coal production in India totaled 610 Mt in 2013, and 668 Mt in 2014. CIL is a 90% SOE (cf. https://www.coalindia.in/en-us/company/structure.aspx). 116 According to IEA (2015c), Indonesia has produced 487.7 Mt of coal (total of thermal coal and metallurgic coal) in 2013.
204 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
“stranded cost compensation”. However, this compensation is paid to electricity producers to compensate for the termination of long-term power-purchase agreements. In context of the present analysis, this does not constitute a producer subsidy but rather a consumer subsidy as power plants are the consumers of the coal and such subsidies have been excluded in other cases as well. A small fraction of 0.01 bn USD equivalent to a subsidy of 0.1 USD/t remains, which is attributed to free energy supply for mine workers.
Russia
Data on fossil fuel subsidies is available from Ogarenko et al. (2015) and OECD (2015a). For coal production estimates diverge very significantly, with 0.07 bn USD estimated by the former for 2013- 2014, while the latter reports an average subsidy level of 0.99 bn USD for 2006-2014, and an extreme increase to levels of 2.29 bn USD to 6.04 bn USD for 2013 and 2014. The two sources report very different subsidy levels for the cost items of “Spending on Exploration and Prospecting for Coal” and another item, “Support for Restructuring and Development of the Coal”, is missing in the former database. For reasons of consistency, I base the analysis on data from Ogarenko et al. (2015). Accounting for average annual steam coal production in 2013 and 2014, the resulting subsidy level is 0.4 USD/t. The majority of subsidies is directed towards tax benefits on the regional level (52%), while the rest is used for tax exemptions (33%) and direct spending (14%).
South Africa
Information on coal production subsidies is rarely available for South Africa. Garg and Kitson (2015) report expenditures of 0.04 bn USD to expand coal transportation infrastructure in 2014.
Beside, Eberhard (2015, 180) states that Sasol sells underpriced coal to its Coal-to-Liquids (CtL) plant at rates of 12 USD/t while domestic coal prices are reported to be 20 USD/t (Eberhard 2015, 196). In the COALMOD-World base case data, production cost for South Africa do not account for this subsidy and start at 20 USD/t. Therefore, only subsidies to transport infrastructure which is dedicated to export coal are included in the new analysis, which accounts for 0.5 USD/t.
USA
Data on current coal production subsidies in the US is available from different sources: the reported levels range from 1bn USD in 2013 reported by EIA (2015a) to 6.8bn USD calculated by Fulton, Buckley et al. (2015). Doukas and Whitley (2015) undertake an extensive review of fossil fuel related subsidies in the US and provide a detailed list by subsidy type, jurisdiction, fuel, and fuel chain stage (ODI 2015c) which extends the effort undertaken by OECD (2015a). They calculate national coal production subsidies of 2.1 bn USD. According to their figures, the largest shares of total subsidies originate from relief of royalties (50%), and support for extraction (15%), and remediation (18%). The data allows differentiating between federal subsidies, which apply to all US coal production (44%), and state subsidies that only apply to particular basins (subsidies by the state of Wyoming make up 50% of
205 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
total subsidy level). Based on this disaggregation, coal production subsidies for the Powder River Basin amount to 3.4USD/t, 1.1USD/t for Appalachia and 1.0USD/t for all other basins117.
Other producers
No information on subsidies is available for Colombia, and other smaller coal producers like Venezuela and Mozambique.
A.8 Country-by-country assessment of coal reserves in operating mines Australia
Values for Australia are based on information from Geosience Australia (Britt et al. 2015) which report 19816 Mt reserves of black coal in operating mines. Mine level information is available from Australian Mines Atlas118 but could not be used due to a lack of reporting by some companies.
China
The Statistical Yearbook 2014 (NBSC 2016, Table 8.5) reports ensured reserves by region for 2013, which sum up to a total of 236290 Mt. To arrive at a number on recoverable reserves the figure is corrected by the average recovery factor of 48% obtained from Zhang et al. (2016). Data on reserves in producing mines could not be obtained, as there seems to be no obligation to publish such information to the general public. Therefore, numbers are calculated based on the ratio of reserves reported by BGR (2015) to reserve in operating mines directly obtained from literature (for USA, Colombia, Poland, South Africa, Indonesia, and Australia). The number in brackets is based on the highest ratio obtained in South Africa (69%), while the standard assumption is the average ratio (33%).
Colombia
For Colombia, data on coal reserves in operating mines is obtained from the annual reports of the operating companies which were available for Cerrejon, Calenturitas, and La Jagua from Glencore (2016, 59). The operator Caribbean Resources Corp. provides only resource estimates for its mines Cerro Largo (11.6-21.2 Mt), and La Caypa (47 Mt)119. The operator Drummond Company does not provide any data on reserves, instead estimates from the Global Methane Initiative (EPA 2015a, 85) where used for La Loma (485 Mt), and El Descanso (960 Mt). No data was found for La Francia and Jam. In total, 3221 Mt of reserves are estimated in operating mines in Colombia.
India
117 According to EIA (2016b), US coal production totaled 984 Mt (short) in 2013 and 1000 Mt (short) in 2014; Wyoming had an average share of 39%, West Virginia (11%) and Kentucky (8%). Federal level subsidies that apply to all production sites account for 44% of the totals, while West Virginia and Kentucky account for 4% and 2%, respectively; the remained stems from Wyoming state support. Furthermore, the calculation takes into account the share of Wyoming of total PRB production (86%) and the share of Kentucky and West Virginia in total Appalachia production (65%). 118 http://www.australianminesatlas.gov.au/?site=atlas&tool=search. 119 http://www.caribbeanresources.ca/Properties/Map-of-Properties/default.aspx
206 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
On the one hand, information on coal reserves in India is readily available from the Coal Directory of India Coal statistics Controller's Organisation (2015), differentiated by depth, quality, and certainty, on a field-wise level. On the other hand, this data carries high uncertainty and measurement is not in line with international standards. Fernandes and Sanzillo (2013) report that reserve estimates for Coal India Limited (CIL), India’s largest, state-owned coal company are 17% overestimated because of categorization based on India ISP code, instead of the international common UNFC code. Cmpdi (2014) provides estimates of category G1 reserves120 of 19805 Mt of mineable coal in operating mines. This number excludes reserves in captive mining blocks that where allocated to private and public companies. Reserves in these deposits are only available from the Coal Directory of India Coal statistics (2015), reported at 34419 Mt. Applying the same correction factor as for coal from CIL deposits gives a total of 48373 Mt. Currently, the allocation of all these blocks except for four was found illegal and arbitrary by the Indian supreme court in 2014 (Rajagopal 2014). The court ruled that the central government has to re-auction these blocks or has to collect adjustment payments instead. I assume that these blocks remain undistributed as an extreme assumption, resulting in the difference between the high and the low estimate reported for India.
It is worth noting that a significant share (24%) of India’s coal reserves are in low quality coal with an energy content of 4600 Kcal/Kg and below (see Table 2.5, Coal Controller’s Organisation 2015). Assuming that India will pursue modernizing its coal power plant fleet to achieve higher efficiency and lower specific local and global emissions, it will need to rely on higher quality reserves, which would render low quality deposits stranded.
Indonesia
For Indonesia, active mining operations were identified by assessing the Mining Atlas121 with limited free access. A mine-by-mine assessment was performed based on company annual reports and other publicly available data. Some sources do not distinguish between proven and probable reserves. Total reserves in operating mines are estimated to be 3.5 Gt as of end of 2015.
Kazakhstan
For Kazakhstan no data on reserves in operating mines could be found. The World Energy Council (2013) reports recoverable reserve of hard coal of 21.5 Gt assuming the same ratio between estimated reserves and reserves in operating mines as estimated for Ukraine and Russia (Ukraine: 2500/(15351+16577); Russia: 17700/(49088+97472), resulting in on average 10%). Therefore, reserves in active mines are estimated at 2.2 Gt.
Poland
120 Defined as: feasibility study (F1) has been made and economically viable (E1). The balance Mineable Reserve (excluding that of losing mines) as on date will be in this category (cmpdi 2014, 9). 121 https://mining-atlas.com/operation/php.
207 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Saboczyk and Salagua (2013) compare reserve estimates obtained from using the Polish methodology for reserve assessment to estimates based on the JORC code. They report 0.8 Gt of coal in operating mines under valid concessions.
Russia
According to SUEK (2011), the largest Russian coal producers, the company’s proven and probable reserves in operating mines totaled 5.9 Gt by April 2011. The company accounts for 33% of coal production in Russia in 2011 (Tazazanov 2012). For other major Russian coal companies no data on reserves was publicly available. To estimate total reserves, I assume that reserves are evenly distributed among the Russian coal mining companies; therefore I assume that SUEK holds 33% of coal reserves in operating mines. This gives a total estimate of 17.7 Gt.
South Africa
Using data from Wood Mackenzie, SACRM reports reserves of operating mines at 8.9 Gt in 2010, with 95% of this being thermal coal and the remainder metallurgical in operating mines (SACRM 2011): Low initial coal quality requires washing and beneficiation before coal can be marketed, and 21-24% of initial mined run-of-mine coal is discarded (SACRM 2011). Accounting for discard, the remaining steam coal reserves in operating mines are estimated at 6.8 Gt.
Ukraine
Due to a lack of available reserve data from the majority of coal mining companies, a similar approach as in the case of Russia is chosen for Ukraine. DTEK reports commercial coal reserves of 1.7 Gt as of 01.01.2015 (DTEK 2014). The company currently produced 46 Mt of coal in 2014, which accounts for 69% of total production in Ukraine reported by DTEK. Numbers on annual production might be significantly reduced, due to the armed conflict in Ukraine, especially in the largest coal production region Donetsk. Due to a lack of other sources, this number is used to scale up reserve figures to 2.5 Gt.
USA
Figures on estimated recoverable reserves are reported as 232017 Mt (255755 short Mt) in EIA (2016b, Table 15). Figures on reserves in operating mines are given as 17555 Mt (19351 short Mt) in EIA (2016b) based on data from EIA (2016b, Table 14).
Other small producers
For other, small producers, the original entry for estimated reserves from the COALMOD-World data base (Table A.6 in Appendix A) is used, as they will not have a major influence on international trade patterns, prices and emissions, due to their insignificant size.
208 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.9 Further results of scenario with high estimate of reserves in operating mines Table A.11: Cumulative production in reference case and scenario with high estimate of reserves in operating mines (in Mt). Cumulative production [Mt] Cumulative production [Mt] Reference Reference M&E Country M&E Scenario Change in Country Change in case case Scenario AUS 6.4 8.5 33 POL 1.9 0.8 -58 CHN 89.7 83.7 -7 RUS 6.6 10.0 52 COL 4.9 3.2 -35 UKR 1.5 2.1 40 IDN 13.0 6.1 -53 USA 32.3 14.9 -54 IND 30.1 26.9 -11 VEN 0.5 0.5 0 KAZ 3.5 2.2 -37 VNM 0.2 0.2 0 MNG 1.2 1.2 0 ZAF 12.0 6.8 -43 MOZ 0.2 0.2 0 Total 204.0 167.3 -18
Figure A.7: Total supply from imports and domestic production in scenario with high estimate of reserves in operating mines (in Mtpa).
209 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.10 Further details and results of the M&E scenario Table A.12: Cumulative production in reference case and M&E scenario (in Mt). Cumulative production Cumulative production [Mt]
[Mt] Reference M&E Change in Reference M&E Change in Country Country case Scenario % case Scenario % AUS 6447 4000 -38 POL 1916 2505 31 CHN 89723 50113 -44 RUS 6645 3970 -40 COL 4865 3521 -28 UKR 1455 940 -35 IDN 13000 11850 -9 USA 32291 13001 -60 IND 30146 10916 -64 VEN 479 479 0 KAZ 3481 2090 -40 VNM 150 150 0 MNG 1170 1170 0 ZAF 12033 2788 -77 MOZ 212 212 0 Total 204013 107705 -47
Figure A.8: Total supply from imports and domestic production in M&E scenario (in Mtpa).
210 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
A.11 Conversion between reserves and resources data from McGlade and Ekins (2015) and COALMOD-World There is no perfect match between the data provided by McGlade and Ekins (2015) (M&E) and data from the COALMOD-World (CMW) dataset. While TIAM-UCL, the model used by M&E, is an energy systems model, CMW is a sectoral model. Additionally, the two models have different time horizons: TIAM-UCL calculates energy use until 2100, with levels of resources unburned reported until 2050.
The two models also have different spatial coverage. While the first has a global coverage of both the demand and the supply of fossil resources, the second is focused on international trade aspects and therefore has no representation of purely self-supplying countries. Therefore, both supply and demand estimates are lower in COALMOD-World as compared to TIAM-UCL.
Moreover, M&E’s definition of hard coal also includes coking coal, while CMW is focused on steam coal only, which increases hard coal demand in M&E compared to CMW. Additionally, M&E overestimates the use of lignite compared to hard coal (cf. BGR (2015): hard coal production in 2014: 7.15 Gt, Lignite: 1.05 Gt, compared to 4.9 Gt hard coal and 3.6 Gt lignite calculated by M&E). One explanation for this divergence might be that M&E overestimate the mobility of lignite which is only used for local electricity production due to its low energy content per volume and tonnage ratio, in reality. This might lead to an overestimation of unburned hard coal reserves in M&E. To be consistent with other literature (e.g., Finighan 2016), I use estimates on shares of unburned coal reserves, which comprise hard coal and lignite, for my scenario calculations.
For neither of the issues discussed above there is an easy fix or work-around. As many other measurement errors are also inherent to both model datasets, the results of the scenario should be interpreted as approximate values.
211 Appendix A Mathematical Formulations and Additional Data for Chapter 2, 3, 4
Table A.13: Production capacity and reserves in COALMOD-World dataset and M&E scenario. Production M&E Scenario Cap. in Reserves in values for coal COALMOD- COALMOD- reserves burned World World Scenario until 2050 Comments USA P_USA_PRB 525 112555 6469 12 Gt for USA Distribution based on current production levels P_USA_Rocky 79 20704 973 same as above same as above P_USA_ILL 115 82887 1419 same as above same as above P_USA_APP 336 54572 4140 same as above same as above Colombia P_COL 75 6229 3521 4 Gt for CSA Distribution ensures usage of Venezuela reserves Venezuela P_VEN 10 479 479 same as above same as above Poland P_POL 71 13997 9000 9 Gt for Europe Ukraine P_UKR 45 16271 940 7 Gt for FSU Distribution based on current production levels Kazakhstan P_KAZ 100 28145 2090 same as above same as above Russia P_RUS 190 49078 3970 same as above same as above South Africa P_ZAF 267 48740 2788 3 Gt for Africa Distribution ensures usage of Mozambique reserves India P_IND_North 281 35663 6169 62 for India and Distribution based on current reserves and ensures usage of China Mongolia reserves P_IND_Orissa 123 14416 2494 same as above same as above P_IND_West 53 7134 1234 same as above same as above P_IND_South 58 6755 1169 same as above same as above Vietnam P_VNM 62 150 150 12 Gt for Other Distribution ensures usage of Vietnam reserves developing Asia Indonesia P_IDN 340 13000 11850 same as above same as above China P_CHN_SIS 1573 213400 36916 62 for India and Distribution based on current reserves and ensures usage of China Mongolia reserves P_CHN_Northeast 121 15900 2750 same as above same as above P_CHN_HSA 564 4700 4700 used up until 2020, before policy is introduced P_CHN_YG 450 36800 6366 same as above same as above Australia P_AUS_QLD 85 24764 1667 4 for OECD Pacific Distribution based on current production capacity P_AUS_NSW 119 13829 2333 same as above same as above Mongolia P_MNG 17 1170 202 62 for India and Distribution based on current reserves and ensures usage of China Mongolia reserves Mozambique P_MOZ 5 212 212 3 Gt for Africa Distribution ensures usage of Mozambique reserves Source: Own calculation based on reserves and production data from Holz et al. (2015), and McGlade and Ekins (2015).
212 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
Appendix B MATHEMATICAL FORMULATIONS AND ADDITIONAL DATA FOR CHAPTER 5, 6,7
B.1 CCTS-Mod: additional data and results Table B.1: Definition of indices, parameters, and variables of CCTS-Mod Description
Sets a,b Model period D Pipeline diameter [m] i,j Node P Individual CO2 producer S Individual CO2 storage site
Parameters c_ccsPa Variable costs of CO2 capture for producer P [€/t CO2 per year] c_f CO2 flow costs [€/t CO2 per year] c_inv_fd Pipeline investment costs [€/km*m (diameter)] c_inv_xP Investment costs of CO2 capture for producer P [€/t CO2 per year] c_inv_ySa Investment costs for storage in sink S [€/t CO2 per year] c_plan Pipeline planning and development costs [€/km] cap_dd Capacity of a pipeline with diameter d [t CO2/a] cap_stor Storage capacity of sink S [t CO2] capt_rate Capture rate for CO2 capture [in these scenarios: 90%] certa CO2 certificate price [€/ t CO2] CO2Pa Total annual quantity of CO2 produced by producer P [t CO2] Eij Distance matrix of possible connections between nodes i and j match_PPj Mapping of producer P to node j {0;1} match_SSj Mapping of sink S to node j {0;1} max_pipe Maximum number of pipelines built along planned route r Rate of interest [%] start Starting year of the model yeara Starting year of the model period a
Variables fija CO2 flow from node i to j [t CO2/a] inv_fijda Investment in additional pipeline capacity with diameter d inv_xPa Investment in additional CO2 capture capacity from producer P [t CO2/a] inv_ySa Investment in additional injection capacity of sink S [t CO2/a] planija Pipeline planning and development between nodes i and j xPa Quantity of CO2 captured by producer P [t CO2/a] ySa Quantity of CO2 stored per year in sink S [t CO2/a] zPa Quantity of unabated CO2 emitted into the atmosphere [t CO2/a]
Table B.2: Estimated CO2 storage potential
213 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
Saline Depleted Offshore Offshore Aquifer Gasfield Aquifer Gasfield Total Country [GT CO2] [GT CO2] [GT CO2] [GT CO2] [GT CO2] Austria 2.30 2.30 Belgium 0.30 0.30 Bulgaria 1.70 1.70 Bosnia and Herzegovina 0.20 0.20 Czech Republic 0.70 0.70 Germany 3.80 1.60 1.20 6.60 Denmark 2.50 2.50 Spain 11.00 3.50 14.50 France 5.70 5.70 Greece 0.30 0.30 Croatia 2.80 2.80 Hungary 0.20 0.20 Ireland 2.00 1.30 3.30 Italy 5.50 5.50 Latvia 1.30 1.30 Macedonia 0.30 0.30 The Netherlands 0.70 0.50 1.20 Norway 1.90 11.90 13.80 Poland 3.70 0.70 3.50 7.90 Romania 0.40 0.40 United Kingdom 14.40 7.80 22.20 Total 40.90 3.00 29.60 20.20 93.70 Source: Own calculations based on various studies (GeoCapacity 2009; Brook et al. 2009; M. Bentham 2006; M. S. Bentham, Kirk, and Wiliams 2008; Radoslaw, Barbara, and Adam 2009; Greenpeace 2011, 20; Hazeldine 2009; Ainger, Argent, and Haszeldine 2010).
214 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
Figure B.9: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, On50
Figure B.10: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, On100
Figure B.11: Storage by sector in MtCO2 and infrastructure investment and variable costs in €bn, Off100
215 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
Table B.3: Summary of scenarios results and CO2 price assumptions for Chapter 6 Scenario Pipeline Network Number of Stored Origin. Storag CCTS CCTS [1000s km] wells* Emiss. until from e left invest. var. [GtCO2] industr in costs costs 2030 2050 2030 2050 2030 2050 y [%] 2050 [€bn] [€bn] [GtCO 2] EU_40% - <1 0 2 - 0.02 100 50.0 0.2 0.4 EU_80% 1.4 45 34 1176 - 12.2 55 37.9 306.6 731.2 NorthSea 14.2 15.4 119 141 0.6 2.5 100 40.0 47.2 150.0 _40% NorthSe 10.2 26.8 122 760 0.6 8.5 54 34.6 191.9 539.3 a_80% DNNU 11.0 13.6 110 174 0.6 3.1 57 36.4 61.7 232.4 _80% * annual well capacity for CO2-EOR: 1MtCO2; DOGF:0.8 MtCO2 ; Saline Aquifers: 0.8 MtCO2 following (IEAGHG and ZEP 2011) and own assumptions.
Scenario 2015 2020 2025 2030 2035 2040 2045 2050 Certificate 40% 14 17 27 37 45 52 52 52 price in 80% 18 25 39 53 75 97 183 270 €/tCO2
216 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
B.2 Combined Electricity and CCTS Investment and Dispatch Model (ELCO): Karush-Kuhn-Tucker conditions
B.2.1 The electricity sector LTN, : gh,,, n t a
mu_ ehna,, EF_ ELt 1 CR _ G t CPSa EUA a (B.1) DFah PDa TD VC__ G INTC G ·g 0 n,, t a t h,,, n t a gh,,,n t a 0 target_2 CO · a ta, emps cap__ g curt el TDh n,,,,,, t a EF_1 EL t CG R _ t h n t a h a
LTN, : g _ cfdh,,,, n t aa a target_2 co SPt,, aa t aaa·aaa aaa I_ USEt,, aaaa a target_ RE 1TARGET _ REaaa · aaa aaa I__, USE ELt,, aa aaa t T_ RES target_ RE 0 DF PD TD TARGET_· REaaa aaa a a h aaa I__, USE ELt, aa, aaa (B.2) tT _ RES EF_ EL 1 CR _ G CPS EUA t t a a EF_ EL ·CR _ G · mu _ co 2 t t h,, n a VC__ Gn,, t a INTC Gt ·_g cfdh,,, n t, aa a emps cap___ g cfd curt el TDh n,, tt a EF_ EL t 1 CR _ G t h,, n,,, t aa a h a tt ONEFUELtt, t diff___ g diff g diff g TDht, a TD h DIFF_ G t· t , a 1 t ,a 2 g_0 cfd h,,, n, t aa a
LTN, : inv_ gn,, t a PDaa DFaa FC__ G n,,,, t aa INVC G n t aa aaI _USE_ ELt,, a aa emps TDh· AVAILh,,,, n t EMPS a n tt aa haa I__ USE ELt,, a aa tt ONEFUELtt, t inv _ g 0 h,,, n t a (B.3) cap_ g 0 AVAILh,,,, n t h n t, aa h aaI__ USE EL t,, a aa cap__ g cfd AVAILh,, n t ·hn, ,,,t a aa h aaUSE_ ELt,, a aa pot_ g n, t, aa aaI__ USE ELt,, a aa
217 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
LTN, emps : n,, t a
TDh· AVAILh,,,, n t inv_ g n tt aa EMPS aa h aa USE_, ELt,, a aa (,)_t tt ONE FUELt, tt (B.4) g EF_ EL 1 CR _ G emps 0 h,,, n t a t t n,, t a 0 TDh· g_ cfdh,,,, n tt aa a EF _ EL tt 1 CR _ G tt h aa USE_, ELt,, a aa t,_ tt ONE FUELt, tt
LTN, cap_ g : h,,, n t a 0AVAIL INICAP _ G inv _ g g cap_ g 0 (B.5) hnt,,,,,,,,,,,, nta ntaa hnta hnta aaUSE_ ELt,, a aa
LTN, cap__ g cfd : h,,,, n t aa a (B.6) cap__ g cfd 0AVAILhnt,,,,,,,,,,,, inv _ g ntaa g _ cfd hntaaa hntaaa 0
LTN, : pot_ g n,, t a (B.7) pot_ g 0MAX _ INVn,,,, t inv _ g n t aa n t, a 0 aaUSE_ ELt,, a aa
LTN, diff_ g : ta,
AVAILh,, n t·TD h hn, 0 START___ Gt · TDh · g cfd h, n ,t , aa , a 1 g cfd h , n , t ,aa a,2 · DIFF_ G t (B.8) #of nodes h,, n aa diff_ g TDh· g_ cfdh,,,,n t aa a t, a 0 hna, , a
B.2.2 Shared environmental constraints for the electricity sector
0PD· TD g g _ cfd target_2 co 0 (B.9) athhnta,,,,,,, hntaaa,a a h, n,_ t aa USE ELt,, a aa
218 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
g__ cfdh,,,,,, n t aa a RES OLD h n a aa USE_, ELt,, a aa tT _ RES target_ RE 0PDah··TD a 0 hn, RE_ TARGET d a h,, n a hn, (B.10)
B.2.3 The electricity transportation utility LTSO_ E : el_ t cap_ el 0DFa PDa TD h mu _ el hna,,,,,,,,,,, mu _ el hnna VC __ EL T nnn hnnna el_ t hnnan 0 (B.11)
LTSO_ E : inv__ el t cap____ el t cap el t 0PDaa DFaa INVC __ EL T n,,,,,,,,,,, nn ADJ _ EL n nn h n nn aa h nn n aa inv__0 el t h n nn a aa a h aa a
(B.12)
LTSO_ E cap__ el t : h,,, n nn a
0INICAP _ EL _ Tn,,,,,,,,,, nn ADJ _ EL n nn inv _ el _ t n nn aa ADJ _ EL nn n inv _ el _ t nn n aa el _ t h n nn a aa a cap__ el t h,,, n nn a 0
(B.13)
B.2.4 The industry sector
LIN, : co2_ c h,,, n i a co 2 _ ch,,, n i a 0 (B.14) max_ ind cap _ co 2 _ c 0DFa PD a TD h EUA a mu _ co 2 hna,,,,,,,,,, VC _ CO 2 nia hnia hnia
219 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
LIN, : inv_ co 2 _ cn,, i a PDaa DF aa FC_ CO 2 n,,,, i aa INVC _ CO 2 n i aa aaI_ USE _ CO 2i,, a aa 0 cap_ co 2 _ c CR _ IND inv _ co2 _ c 0 (B.15) h,,, n i aa i n,, i a h aaI_ USE _ CO 2i,, a aa diff_ co 2 _ c diff _ co 2 _ c i,, a i aa DIFF_2 CO i aa a
LIN, max_ ind : h,,, n i a (B.16) max_ ind 0CO 2 _ INDhnia,,,,,,,,, CR _ IND i co 2 _ c hnia hnia 0
LIN, : cap_ co 2 _ c h,,, n i a (B.17) cap_ co 2 _ c inv_ co 2 _ cniaa,, CR _ IND i co 2 _ c hnia,,,,, hni, a 0 aaUSE_ CO2i,, a aa
LIN, : diff_ co 2 _ c ia, (B.18) diff_ co 2 _ c 0START _2 COi inv _2_ co c niaa,,,,, DIFF_2 CO i inv _2_co c nia ia 0 n aa a n
B.2.5 The CO2 transportation utility LTSO_2 CO : co2_ th,,,n nn a (B.19) cap_ co 2 _ t 0DFa PDa TD h mu _2 co hnna,,,,,,,,,,, mu _2 co hna VC _2_ CO t nnn hnnn ah co2_ t n nn a 0
LTSO_ E : inv_ co 2 _ t cap_ co 2 _ t cap _ co 2 _ t 0PDaa DFaa INVC _ CO 2 _ T n,,,,,,,, nn ADJ _ CO 2 n nn h n nn aa h nn n aa (B.20) aaa h aa a
inv _ co 2 _ th,,, n nn a 0
220 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
LTSO_ E cap_ co 2 _ t : h,,, n nn a
0INICAP _2_ CO Tn,,,,,,, nn ADJ _2 CO n nn inv _2_ co t n nn aa ADJ _2 CO nn n inv _2_ co t nn n aa aa a cap_ co 2 _ t co2_ th,,,,,, n nn a h n nn a 0 (B.21)
B.2.6 The CO2 storage sector LSN, : co2_ sh,,, n s a
EFF_2 CO OILPRICEa DF PD TD co 2 _ sh,,, n s a 0 a a h mu_ co 2 VC _ CO 2 INTC _ S · co 2 _ s 0 hna,,,,,,, nsa t hnsa TD· PD max_ stor cap _ co 2 _ s hhaa n,,,,, s aa h n s a hh aa a (B.22)
LSN, : inv_ co 2 _ sn,, s a PDaa DF aa FC_ CO 2 n,, s aa INVC _ CO 2 n,, s aa aaI_ USE _ CO 2s,, a aa 0 inv _ co 2 _ s 0 (B.23) cap_ co 2 _ s diff _ co 2 _ s diff _ co 2 _ s n,, s a hnsaa,,,,, sa saa DIFF_2 CO s h aaI__ USE CO2s,, a aa aaa
LSN, : cap_ co 2 _ s h, n ,s, a (B.24) cap_ co 2 _ s 0 inv _ co 2 _ sn,s, aa co 2 _ s h , n ,s, a h , n ,s, a 0 aaUSE_ CO2s,, a aa
LS,N : max_ stor n,, s a (B.25) max_ stor 0MAX _ STORns,,,,,, TD h PD aa co 2 _ s hnsaa nsa 0 h aa a
LSN, : diff_ co 2 _ s sa, (B.26) diff_ co 2 _ s 0START _2 COs inv _2_ co s nsaa,,,,, DIFF_2 CO s inv _2_co s nsa sa 0 n aa a n
221 Appendix B Mathematical Formulations and Additional Data for Chapter 5, 6,7
B.2.7 Market clearing conditions across all sectors 0g g _ cfd el _ t el _ t D RES _ OLD hnta,,,,,,,,,,,,,,,,, hntaaa hnnna hnnna hna hna t aa USE_ ELt,, a aa nn nn (B.27)
mu _ ehna,, ( free ) h , n , a
LTN, : curt_ el ha, (B.28) 0D RES _ OLD g g _ cfd curt_ el 0 hna,,,,,,,,,,,, hna hnta hntaaa ha n n,_ t aa USE ELt,, a aa
g___ cfd EF EL CR G h,,,, n t aa a t t t aa USE_ ELt,, a aa 0 co 2 _ c mu _ co 2 ( free ) h , n , a h,,,,, n i a h n a (B.29) i co2 _ t co 2 _ t co 2 _ s hnnna,,,,,,,,, hnnna hnsa nn nn s
222