碳定价与二氧化碳排放弹性（英文版）--未来资源.pdf

Carbon Pricing and the Elasticity of CO 2 Emissions Ryan Rafaty, Geoffroy Dolphin, and Felix Pretis Working Paper 21-33 October 2021Resources for the Future i About the Authors Ryan Rafaty is a political scientist by training, whose ongoing research investigates the design, sequencing, and comparative performance of climate change mitigation policies under varied political and economic regimes. Ryan is a Postdoctoral Researcher at Climate Econometrics, based at Nuffield College, Oxford, where he contributes to policy-focused knowledge exchange between climatology, econometrics, and political science. He has a doctorate from the Department of Politics at University of Cambridge, and is a Fellow at the Cambridge Centre for Environment, Energy and Natural Resource Governance. Geoffroy Dolphin is a postdoctoral fellow at RFF and PhD graduate from the University of Cambridge Judge Business School, where he is an affiliate of the Energy Policy Research Group. Geoffroy’s research interests span political economy, energy you must give appropriate credit, provide a link to the license, and indicate if changes were made, and you may not apply additional restrictions. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. You may not use the material for commercial purposes. If you remix, transform, or build upon the material, you may not distribute the modified material. For more information, visit https:/ /creativecommons.org/licenses/by-nc-nd/4.0/. Carbon Pricing and the Elasticity of CO 2 Emissions Ryan Rafaty, ∗ Geoffroy Dolphin ∗ ∗ and Felix Pretis *** Manuscript Version: October 19, 2021 Word Count: 16,022 (excluding appendices and references) ABSTRACT We study the impacts of carbon pricing on CO2 emissions across five sectors for a panel of 39 countries covering 1990–2016. Constructing new sector-level carbon price data, we implement a novel approach to estimate the changes in CO2 emissions associated with (i) the introduction of carbon pricing regardless of the price level, (ii) the elasticity of emissions with respect to the price level, and (iii) the potential response of future emissions to possible carbon price trajectories. Using a synthetic control factor model, we find that the introduction of carbon pricing has reduced growth in total aggregate (national) CO2 emissions by 1–2 percent on average relative to imputed counterfactuals, with most abatement occurring in the electricity and heat sector. Exploiting variation in observed carbon prices to explain heterogeneity in treatment effects, we decompose the average treatment effect obtained from the synthetic control factor model to distinguish the ∗ Climate Econometrics at Nuffield College, University of Oxford and Institute for New Economic Thinking at the Oxford Martin School ∗ ∗ Resources for the Future and Energy Policy Research Group, University of Cambridge *** Department of Economics, University of Victoria, Climate Econometrics at Nuffield College, University of Oxford, and Institute for New Economic Thinking at the Oxford Martin School 2 effect of merely introducing a carbon price from the effect of the price level itself. We find a small and imprecisely estimated semielasticity of a 0.03 percent reduction in emissions growth per average $1/metric ton of CO2. Simulating the response of future global emissions to several possible carbon price trajectories, we conclude that carbon pricing alone, even if implemented globally at a level equivalent to the world’s current highest recorded price in Sweden, is unlikely to be sufficient to achieve emission reductions consistent with the Paris climate agreement. Keywords: Carbon Pricing, CO2 Emissions Elasticity, Carbon Tax Effects, Emissions Trading Effects, Climate Policy Impact Evaluation, Generalized Synthetic Control, Emissions-Weighted Carbon Price. JEL Classifications: Q43, Q48, Q54, Q58, H23. 3 1. INTRODUCTION Pricing carbon dioxide (CO2) emissions—via a carbon tax, emissions trading system, or some hybrid scheme—has long been recommended as an integral and, in principle, cost-efficient way to reduce emissions and mitigate the adverse impacts of climate change (Baumol and Oates 1988; Nordhaus 1992; Metcalf 2009; Cramton et al. 2017; Stern-Stiglitz High-Level Commission on Carbon Prices 2017). 1 Since the world’s first carbon taxes were implemented in Finland and Poland in 1990, an additional 28 jurisdictions have adopted them. Similarly, since the European Union established the world’s first emissions trading system (ETS) covering CO2 emissions in 2005, the number of carbon markets has grown to 31, with the latest additions in China, the United Kingdom, and Germany in 2021. Carbon pricing initiatives now cover one-fifth of global greenhouse gas (GHG) emissions, or 12 gigatons (Gt) of CO2 equivalent emissions annually. These initiatives raised public revenues totaling US$53 billion in 2020 (World Bank 2021). However, behind the proliferation and popularization of the carbon pricing paradigm is a great uncertainty over its role in climate policy. Critics and endorsers alike concede that “optimal” pricing schemes that are cost-efficient and environmentally effective in theory may be politically unfeasible in practice (Rosenbloom et al. 2020a; Stiglitz 2019). A clash of paradigms persists regarding what this means in practical political terms (Rosenbloom et al. 2020b; van den Bergh 1 The optimal carbon price is typically defined in relation to an ideal objective function that sets the carbon tax rate equal to the monetized damages associated with emitting an additional ton of CO 2 , referred to as the “social cost of carbon” (SCC) (Gillingham and Stock 2018). However, global SCC estimates can be US$10/tCO 2 to US$1,000/tCO 2 and above due to the uncertainties inherent in damage function estimation and alternative ethical parameters (Adler 2017). For policymakers seeking guidance in setting the optimal price level, the unwieldy range of SCC estimates is unhelpful. This has prompted some economic policymakers to advance a target-based approach, whereby the appropriate price path is one that minimizes the cost of achieving a desired quantity of CO 2 reductions over a given period (Hepburn 2017). 4 and Botzen 2020). Under the 2015 Paris Agreement, 195 countries committed to mitigate against dangerous levels of climate change this century by maintaining global average surface temperatures below 1.5–2°C relative to preindustrial conditions, but this would necessitate a reduction in global emissions of around 50 percent by 2030 relative to 2020 (UNEP 2019). 2 In the Economists’ Statement on Carbon Dividends (2019), which claims to be the largest public statement in the history of the economics profession, carbon pricing is hailed as the tool of choice to achieve these reductions at the “scale and speed that is necessary”. 3 According to the Stern-Stiglitz High-Level Commission on Carbon Prices (2017), explicit carbon prices in the range of ≥US$40–80/tCO2 by 2020 and ≥US$50–100/tCO2 by 2030 will be “indispensable” to achieving the Paris Agreement goals, albeit with the proviso that they are combined appropriately with complementary policies. 4 However, such assessments have relied on ex ante calibrated model projections with limited empirical corroboration. For context, current carbon prices range from $1/tCO2 in Poland and Ukraine to $137/tCO2 in Sweden (in nominal terms), and nearly half of all covered emissions worldwide are priced at less than $10/tCO2 (World Bank 2021). 5 Globally, the average (emissions- weighted) carbon price is around $3/tCO2 (Dolphin et al. 2020), equivalent to adding approximately US$0.03 per gallon of gasoline (€0.009 per liter of petrol). 2 This is a necessary but insufficient condition. A further requirement is that global emissions decline to net zero by around 2050–2070. Any irreducible positive emissions would need to be offset by a range of negative emissions technologies, none of which are a panacea and all of which face considerable biophysical limits, uncertain long-term costs, and political coordination challenges (Griscom et al. 2017; Hepburn et al. 2019; Chatterjee and Huang 2020; Smith et al. 2016). 3 The statement (2019) includes among its signatories 3,589 US-based economists, four former chairs of the Federal Reserve, 27 Nobel Laureate economists, and 15 former chairs of the Council of Economic Advisers. 4 As Stiglitz (2019) cautions, carbon price paths will inevitably vary across heterogeneous sociopolitical and economic contexts and, critically, “there is no presumption that a carbon tax alone can suffice to address optimally the problem of climate change” (emphasis in original). 5 As of May 2020. All monetary units throughout this study are in 2015 US dollars. 5 Empirical evaluations of the impact of implemented carbon prices on CO2 emissions have been mixed, inconclusive, and, until recently, strikingly scarce. We report the main empirical findings and evaluation methods of previous studies in Section 3. Our key takeaway from this burgeoning evaluation literature is that the fragmentary nature of the evidence precludes systematic inference about the likely response of emissions to carbon pricing across space and time. As we describe in Section 4, the paucity of cross-country empirical assessments is partly a function of the lack of standardized carbon price data adjusted to account for variation in industry exemptions, rebates, and sectoral coverage. But the empirical neglect can also be attributed to the considerable identification challenges, summarized succinctly by Mildenberger (2020): Carbon pollution levels are so overdetermined by diverse economic and social forces that retrospective causal identification of policy impacts remains difficult. Economists have offered evaluations of some policies, but these estimates are difficult to compare across countries and time. Nor can we reliably translate simple policy content metrics, like a national carbon price level, into units of carbon pollution reduced. Even identical carbon prices have different effects based on variation in sectoral cost exposure and sectoral differences in the elasticity of carbon-dependent activities. Motivated by similar concerns, we present a viable empirical modeling approach that largely overcomes these identification challenges. Until recently, the persistent lack of standardized carbon pricing data has compelled researchers to rely predominantly on quasi-experimental methods to estimate generic “treatment effects” of carbon pricing without specifying the initial price level and its subsequent evolution over the treatment period. In effect, essential information about the dynamics and functional form of the relationship between the price level and emissions is ignored or omitted perforce. This has precluded pursuing conventional economic interest in estimating 6 empirical elasticities (in this case, of emissions, with respect to heterogeneous carbon price levels observed across countries, sectors, and time). Furthermore, when treatment effects or elasticities are estimated, the focus has remained on their statistical rather than economic significance, with few empirically grounded studies assessing whether pricing is sufficient to achieve governments’ emissions reduction commitments. The practical consequence is that policymakers and the public still know little about the environmental effectiveness of one of the core pillars of climate policy. 6 We construct a novel dataset comprising average (emissions-weighted) carbon prices across five sectors for a panel of 39 countries that implemented a carbon price during 1990–2016 (and 164 other countries that did not), combined with emissions data from 1975–2016. We aim to answer three questions. First, does pricing carbon reduce emissions? In other words, what is the effect of the introduction of carbon pricing on CO2 emissions, irrespective of the price level? Second, does the price level matter (do higher carbon prices lead to greater reductions)? Third, is carbon pricing sufficient to achieve international emission-reduction targets? We report two sets of estimated effects for each sector. First, we estimate the average treatment effect of introducing a carbon price irrespective of the price level. To overcome challenges in identifying treatment effects using conventional difference-in-differences (DiD) and synthetic control approaches, we apply treatment evaluation methods accommodating staggered adoption (Xu 2017; Athey et al. 2018) and control for unobserved time-varying heterogeneity using 6 Although much of the academic climate economics discourse has focused on estimating the social cost of carbon (with a view to designing socially optimal carbon pricing schemes), government discourse has shifted toward a more target-based approach since the Paris Agreement. For example, countries accounting for two-thirds of global emissions have announced commitments to achieve “net-zero” CO 2 emissions by midcentury or shortly thereafter, raising the question of whether carbon pricing can plausibly achieve the declared goals and, if not, what role it ought to play in the broader policy mix.