欧洲碳税的宏观经济影响（英文版）The Macroeconomic Impact of Europe’s Carbon Taxes w27488 (1).pdf

NBERWORKINGPAPERSERIES THEMACROECONOMICIMPACTOFEUROPE’SCARBONTAXES GilbertE.Metcalf JamesH.Stock WorkingPaper27488 http://www.nber.org/papers/w27488 NATIONALBUREAUOFECONOMICRESEARCH 1050MassachusettsAvenue Cambridge,MA02138 July2020 WehavereceivedvaluablecommentsfromMeredithFowlieandGeoffroyDolphin.Wewishto thank CelineRamsteinandOzgurBozçagaforhelpwiththeWorldBankcarbontaxdata,and Xiaoxin Zhang for excellent research assistance. The views expressed herein are those of the authorsanddonotnecessarily reflecttheviewsoftheNationalBureauofEconomicResearch. NBERworkingpapersarecirculatedfordiscussionandcommentpurposes.Theyhavenotbeen peerreviewedorbeensubjecttothereviewbytheNBERBoardofDirectorsthataccompanies official NBERpublications. ©2020byGilbertE.MetcalfandJamesH.Stock.Allrightsreserved.Shortsectionsoftext,not to exceedtwoparagraphs,maybequotedwithoutexplicitpermissionprovidedthatfullcredit, including ©notice,isgiventothesource.TheMacroeconomicImpactofEurope’sCarbonTaxes GilbertE.MetcalfandJamesH.Stock NBERWorkingPaperNo.27488 July2020 JELNo.E62,H23,Q43,Q54 ABSTRACT PolicymakersoftenexpressconcernabouttheimpactofcarbontaxesonemploymentandGDP. Focusing onEuropeancountriesthathaveimplementedcarbontaxesoverthepast30years,we estimate the macroeconomic impacts of these taxes on GDP and employment growth rates for variousspecifications andsamples.Ourpointestimatessuggestazerotomodestpositiveimpact onGDPandtotalemployment growthrates.Moreimportantly,wefindnorobustevidenceofa negativeeffectofthetaxonemployment orGDPgrowth.Weexamineevidenceonwhetherthe positive effects might stem from countries that used the carbon tax revenues to reduce other taxes; while the evidence is consistent with this view, it is inconclusive. We also consider the impactofthetaxesonemissionreductionsandfindacumulative reductionontheorderof4to6 percentfora$40/tonCO2taxcovering30%ofemissions.Weargue thatreductionswouldlikely begreaterforabroadbasedU.S.carbontaxsinceEuropeancarbontaxes donotincludeinthe taxbasethosesectorswiththelowestmarginalcostsofcarbonpollutionabatement. GilbertE.Metcalf DepartmentofEconomics TuftsUniversity Medford,MA02155 andNBER gilbert.metcalf@tufts.edu JamesH.Stock DepartmentofEconomics HarvardUniversity LittauerCenterM26 Cambridge,MA02138 andNBER James_Stock@harvard.edu 1 Economists widely agree that putting a price on carbon emissions is the most cost‐ effective way to reduce greenhouse gas emissions. The two most straightforward ways to apply a price are a carbon tax and a cap and trade system. A carbon tax can be levied on fossil fuels and other sources of greenhouse gas emissions based on their emissions; a cap and trade system limits emissions to some overall amount (the cap) and allows polluters to trade the rights to those scarce emission rights. In the current Congress there are numerous bills to establish national carbon tax systems and a few cap and trade bills. The filed bills reflect a growing consensus that action is needed at the national level to curb our carbon pollution and that a carbon tax is the most straightforward way to do that. The bills also reflect a broad consensus among economists, as typified by the more than 3,500 economists who signed the Climate Leadership Council’s statement in calling for a carbon tax as “the most cost‐effective lever to reduce carbon emissions at the scale and speed that is necessary.” 1 A major stumbling block to pricing carbon pollution is concern about the economic impact of the policy. The Trump Administration’s retreat from a climate policy is emblematic. In initiating a process to withdraw the United States from the global Paris Agreement, for example, the President claimed that the cost to the economy would be “close to $3 trillion in lost GDP and 6.5 million industrial jobs…” (Trump, 2017). How should we assess the economic costs of a carbon tax? Until recently, most analyses were based on modeling from large scale computable general equilibrium models. But we now have enough experience with carbon tax systems around the world to carry out statistical 1 The statement was published in The Wall Street Journal on Jan. 17, 2019 and is available at https://clcouncil.org/economists‐statement/. Both of the authors of this paper are signatories of that statement. 2 analyses of existing systems. The first carbon tax was implemented in 1990 so there is now up to three decades of data to draw on. In this paper we carry out an analysis of the 31 countries in Europe that are part of the EU wide emissions trading system (EU‐ETS). While all of these countries price a portion of their emissions through this cap and trade system, fifteen of these countries also impose a carbon tax, mostly on emissions not covered by the EU‐ETS. By limiting our analysis to countries that are part of the EU‐ETS, we can identify the incremental impact of carbon taxes on emissions, output, and employment by leveraging the variation in carbon tax systems within this group of countries. This paper builds on a previous analysis in Metcalf and Stock (2020). We build on that analysis in a number of ways including, importantly, measuring the impact of the carbon taxes on emissions as well as economic growth. We find the following. For a wide range of specifications, we find no evidence of adverse effects on GDP growth or total employment. We also test and generally cannot reject the hypothesis that the carbon tax has no long run effect on growth rates. This finding is consistent with macroeconomic theory that suggests growth rates are driven by fundamentals, such as technological progress, which are unaffected by changes in relative prices. It is also consistent with most general equilibrium modeling of climate policy. Finally, we find cumulative emission reductions on the order of 4 to 6 percent for a tax of $40 per ton of CO 2 covering 30% of emissions. We argue that this is likely to be a lower bound on reductions for a broad‐based carbon tax in the U.S. since European carbon taxes do not include in the tax base those sectors with the lowest marginal costs of carbon pollution abatement. European carbon 3 taxes generally exclude the electricity sector and carbon intensive industries since those emissions are covered under the EU Emission Trading System. The next section provides background and a literature review that places our paper in context. Section III surveys European carbon taxes. Section IV details our data and the econometric approach we take assess the impact of European carbon taxes. Section V presents results from the analysis. The next section presents some robustness results. We provide some concluding remarks in section VII. II. Previous Literature Most analyses of the economic impact of carbon taxes rely on large‐scale computable general equilibrium models. One representative model is the E3 model described in Goulder and Hafstead (2017). They estimate that a $40 per ton carbon tax for the United States starting in 2020 and rising at 5 percent real annually would reduce GDP by just over one percent in 2035 relative to a no‐tax counterfactual. While different models give different results, most find very modest reductions (if at all) in GDP from implementing a carbon tax. 2 Goulder et al. (2019) also consider a U.S. carbon tax starting at $40 per ton and rising at 2 percent annually. They find the GDP costs over the 2016 – 2050 period discounted at 3 percent equal to less than one‐third of one percent of GDP. 2 Trump cited a NERA (2017) study commissioned by an industry group to analyze how meeting an 80 percent reduction by 2050 would affect various industry sectors. Among other issues, the headline number cited by Trump (7 percent reduction in GDP) is from a NERA scenario in which sector specific regulations are imposed with very different marginal abatement costs across sectors. If marginal abatement costs are allowed to equalize across sectors in that study, the costs are reduced by over two‐thirds. 4 Turning to the empirical literature, Metcalf (2019) finds no adverse GDP impact of the British Columbia carbon tax based on a difference‐in‐difference analysis of a panel of Canadian provinces over the time period 1990 – 2016. Using a panel of European countries over the time period 1985 – 2017, he finds, if anything, a modest positive impact on GDP. That imposing a carbon tax might have positive impacts on GDP is not implausible once one considers the governments’ use of carbon tax revenue. In the early 1990s, for example, carbon taxes were imposed in a number of Scandinavian countries as a revenue source to finance reductions in marginal tax rates for their income taxes (see Brannlund and Gren, 1999, for background on these reforms). Variation in the use of revenues from newly enacted carbon taxes could differentially impact economic growth and is something we explore in this paper. Bernard et al. (2018) use a vector autoregression (VAR) to estimate the impact of the BC carbon tax on provincial GDP, controlling for the pre‐tax price of gasoline (or diesel) and US economic variables; they find no impact of the tax on GDP. In earlier work with a more limited version of the data set used in this paper, we ( Metcalf and Stock (2020)) use local projections to estimate the impact of carbon taxes in European countries on GDP and found no adverse impacts of the tax on economic growth or employment. These results are consistent with Yamazaki (2017) analysis of the employment effects of the British Columbia carbon tax. Yamazaki found modest positive impacts on employment in the province. While aggregate impacts were small, he found significant job shifting from carbon intensive to non‐carbon intensive sectors. Focusing on emissions, Lin and Li (2011) estimate difference‐in‐difference regressions comparing individual countries with carbon taxes (Finland, the Netherlands, Norway, Denmark, 5 and Sweden) with a set of control countries and find mixed results. In 4 of the 5 countries, the growth rate of emissions falls by between 0.5 and 1.7. Only the estimate for Finland is statistically significant at the 10 percent level, with the coefficient suggesting a drop in the growth rate of emissions of 1.7 percent. Martin et al. (2014) assess the United Kingdom’s Climate Change Levy’s (CCL) impact on energy and emissions indicators for various manufacturing sectors. As discussed in Metcalf (2019), the CCL is not a true carbon tax given its differential taxation of fossil fuels. While CO 2 emissions fall by 8.4 percent, but imprecisely estimated, their results are also consistent with the CCL leading to fuel substitution away from electricity and toward coal. This follows from the lower tax rate on coal than natural gas. A recent paper by Andersson (2019) focuses on the impact of Sweden’s carbon tax on transportation emissions. He focuses on transportation as this is the sector most impacted by the Swedish carbon tax. He finds an emissions reduction on the order of 11 percent. While this might appear modest given the fact that Sweden has the highest carbon tax in the world, most analysts argue that the transportation sector is the most difficult sector to decarbonize given the efficiency of the internal combustion engine. Turning to British Columbia, Rivers and Schaufele (2015) find that the province’s carbon tax, which covers gasoline, diesel, and natural gas, significantly reduces gasoline consumption. They estimate that the carbon tax has a stronger impact on gasoline demand – by a factor of four – than a comparable increase in the price of gasoline, a surprising finding that the authors attribute to the high salience of the carbon tax. Metcalf (2019) estimated difference‐in‐ difference regressions using Canadian province data and find that the BC tax reduced emissions 6 on the order of 5 to 8 percent since it’s imposition in 2008. Prettis (2019) estimates a 5 percent reduction in transportation emissions from the BC carbon tax, with potentially larger long‐run emissions, but does not detect an economy‐wide emissions reduction attributable to the tax. As noted at the outset, this paper builds on Metcalf and Stock (2020). In addition to considering additional econometric model specifications for employment and GDP, we also assess the carbon taxes impacts on country emissions. We also test whether macroeconomic outcomes are affected by the use of carbon tax revenue. Specifically, we consider whether green tax reforms – reforms where carbon tax revenues are used to lower existing distortionary tax rates – has a different impact on macro outcomes than when the revenue is simply added to general revenue. 3 III. Carbon Taxes in Europe Carbon taxes were first enacted in Europe with Finland leading the way in 1990. Following an early wave of carbon tax enactments primarily in the Nordic countries, more countries enacted carbon taxes and currently sixteen European countries have carbon taxes in place. We focus on the so‐called EU+ countries that are also part of the EU‐ETS and so exclude Ukraine from our analysis. We focus on EU+ countries to control consistently for the impact of the ETS on growth. The ETS went into effect with a pilot phase (Phase I) in 2005. In Phase I, power stations and certain energy intensive sectors were subject to the cap. 4 Phase II (2008 – 3 We can’t rule out the possibility that adding carbon tax revenues to general revenue allows a country to avoid a future tax increase as opposed to an increase in spending. In that case, we would not expect a different outcome than when the revenue is explicitly earmarked for reductions in distortionary tax rates. 4 The sectors are power stations and other combustion plants of at least 20 MW, oil refineries, coke ovens, iron and steel plants, cement clinker, glass, lime, bricks, ceramics, pulp, and paper and board. Aluminum, petrochemicals, ammonia, nitric, adipic, and glyoxylic acid production. and CO 2 capture, transport, and storage were added in Phase III. 7 2012) added domestic aviation (in 2012), and Phase III (2013 – 2020) added various additional sectors. 5 Countries with carbon taxes are as follows (listed in chronological order of enactment): 6 Finland (1990): The first wave of carbon taxes in Europe began with Finland’s enactment of a tax in 1990. As was also the case in subsequent Nordic carbon tax implementations, Finland’s carbon tax was enacted during a time of income tax reform to lower high marginal income tax rates (Carl and Fedor, 2016). Between 1990 and 1994, Finland taxed all fossil fuels including gasoline, diesel, fuel oil, coal, natural gas with some fuels (e.g. n