The European Union’s Emission Trading System (EU ETS) is the main policy instrument to cut greenhouse gas emissions within the EU. This work specifically focusses on the labor market cost effects of the introduction of the EU ETS, separated by the individual effects for the three economic actors: firms, households, and the state. In the form of a literature analysis, the main results for these individual actors are recontrasted and the central research question of who wins and who loses under EU emissions trading is answered.
It could be found that households, in particular low-income groups, are the main losers of the system, since they substantially bear the higher production costs of firms. Firms, on the other hand, although being the direct emitters of greenhouse gases, are not significantly affected by the system through the creation of various compensation mechanisms. Therefore, this analysis classifies them as the winners of the system. Moreover, the state is also considered a winner of the EU ETS, as it directly receives the profits from the auctioning of emission rights. In turn, the redistribution mechanism it chooses largely determines the extent of losses and the costs borne by households.
Table of Contents
1 Introduction
2 Distributional Effects of Cap and Trade Systems on the Labor Market
2.1 Functioning of a Cap and Trade System
2.2 Distributional Effects of a Cap and Trade System
3 Practical Example: Introduction to the EU ETS
4 Literature Review
4.1 Effects on Firms
4.1.1 Context
4.1.2 Methods and Data Samples in the Literature
4.1.3 Evidence: Impact of EU ETS on Employment and Profitability
4.1.4 Critique
4.1.5 Discussion
4.2 Effects on Households
4.2.1 Context
4.2.2 Methods and Data Samples in the Literature
4.2.3 Evidence: Impact of EU ETS on Income and Wages
4.2.4 Critique
4.2.5 Discussion
4.3 Effects on State
4.3.1 Context
4.3.2 Methods and Data Samples in the Literature
4.3.3 Evidence: Impact of EU ETS on GDP
4.3.4 Critique
4.3.5 Discussion
4.4 Summary
5 Policy Implications for the EU ETS and the Labor Market
6 Conclusion
7 References
8 Tables and Figures
Abstract:
The European Union's Emission Trading System (EU ETS) is the main policy instrument to cut greenhouse gas emissions within the EU. This paper specifically focusses on the labor market cost effects of the introduction of the EU ETS, separated by the individual effects for the three economic actors: firms, households, and the state. In the form of a literature analysis, the main results for these individual actors are contrasted and the central research question of who wins and who loses under EU emissions trading is answered. It could be found that households, in particular low-income groups, are the main losers of the system, since they substantially bear the higher production costs of firms. Firms, on the other hand, although being the direct emitters of greenhouse gases, are not significantly affected by the system through the creation of various compensation mechanisms. Therefore, this analysis classifies them as the winners of the system. Moreover, the state is also considered a winner of the EU ETS, as it directly receives the profits from the auctioning of emission rights. In turn, the redistribution mechanism it chooses largely determines the extent of losses and the costs borne by households.
Policy Relevance:
This work contributes to an efficiency-enhancing reform of the EU ETS to promote the shift to an environmental policy that benefits all actors and leaves no one behind.
Keywords: Cap and Trade, CO 2, Emissions Trading System, Labor Market
List of Abbreviations
Abbildung in dieser Leseprobe nicht enthalten
List of Tables
Table 1 The Phases of the EU ETS and their Characteristics
Table 2 Sequential Competitiveness Effects for Regulated F i rm s
Table 3 Literature Overview - Effects on Firms
Table 4 Literature Overview - Effects on Households
Table 5 Literature Overview - Effects on States
Table 6 Summary of Findings
List of Figures
Figure 1 Distributional Effects of a CATS in a Partial Equilibrium
Figure 2 Price Development EU ETS 2006 - 2020
Figure 3 Supply and Demand of Allowances in the EU ETS until 2030
Figure 4 Price development EU ETS 2017 - 2021
Figure 5 Structural Assumptions of the GEM-E3 Approach
Figure 6 Production nests in the GEM-E3 model for the non-energy sector..
Figure 7 Performance of scenario simulations under the GEM-E3 model
1 Introduction
The mitigation of global climate change is one of the greatest policy challenges of the 21st century. The increased accumulation of greenhouse gases (GHG) in the atmosphere caused by human activity, has far-reaching climate-warming and thus environmentally destructive consequences. One instrument that policymakers have developed to reduce global GHG emissions is a so-called Emissions Trading System (ETS). By setting a cap on overall emissions and creating tradable emission allowances equal to the limit set, the system gains control over the total amount of emissions emitted in a certain region. Such cap and trade system (CATS), as the ETSs are also called, has been first established in the European Union in 2005. Since then, it has grown into the world's largest and most strictly regulated system, proven to be an effective tool in emission reduction: Between 2005 and 2019 emissions of covered installations were lowered by about 35 % (European Commission, n.d. a).
However, increasingly stringent emissions reduction targets also come at a cost. Opponents of the EU ETS emphasize that rising carbon prices, combined with an annually decreasing number of emission allowances, are accompanied by large economic costs for all parties involved. One argument that persists in the media is that emissions trading would disrupt labor markets, increases unemployment and thus threatens entire livelihoods (Stratmann, 2021). In particular, the consequences for the labor market are uncertain, as it remains unclear how firms, as the direct participants in the EU ETS, will react to the cost increase of their production factor capital.
This work contributes to the debate on labor market effects of the EU ETS by addressing the individual effects for the three economic actors households (HHs), firms and the state. In form of a comprehensive literature review, covering 51 theoretical ex-ante and empirical ex-post studies, the economic costs borne by these three actors are elaborated. As a main incentive, the question of who are the losers and the winners of the introduction of the EU ETS in relation to the economic costs they bear can be clarified. However, economic benefits in form of environmental improvements will be left out of consideration. With its contribution the paper addresses two existing gaps in the current literature: First, it explicitly focusses on significant evidence for labor market effects of the EU ETS although the evaluation is limited to single selected variables. Second, it comprehensively reviews these effects for all economic actors at the individual level.
This work is structured as follows: First, in a theoretical part, the procedure for analyzing labor market effects through the assessment of distributional effects connected to the introduction of the EU ETS is explained. Subsequently, the analytical general equilibrium model by Fullerton and Heutel (2007, 2010) is introduced, deriving the effect of a climate policy on firms' factor and output prices intuitively. Chapter 3 then provides the reader with more details on the functioning and regulations of the EU ETS. This is followed by the core work of this paper: an extensive literature analysis, divided into the three actors of HHs, firms, and government. For all actors, the same scheme of analysis is applied: First, the considered variables are contextualized and the associated impacts of the EU ETS are presented theoretically. In a next step, the mainly applied methods and data sets found in the literature review are explained, followed by a comprehensive overview of the studies' results. Afterwards, these results are discussed from various points of views, and criticism of existing methodologies and results is given. Based on the literature review, a conclusion can be drawn which actors are the losers and the winners from EU emissions trading. Finally, policy implications for the further procedure with the EU ETS and possible efficiency-enhancing adjustment mechanisms for the EU labor market are derived.
Thus, this work contributes to an understanding of the individual impact of the EU ETS on the main actors of an economy and elaborates possible gaps that need to be improved to create a policy that leaves no actor behind.
2 Distributional Effects of Cap and Trade Systems on the Labor Market
Environmental economists see the pricing of emissions, that accrue as a negative externality in companies' production process, as a fundamental way to reduce the overall emission output in the world's atmosphere. In general, two instruments of carbon pricing can be distinguished: price-based instruments, such as emission taxes and quantity-based instruments like CATSs (Zachmann et al., 2018). In the past, many studies have contrasted the two instruments, emphasizing that CATSs, although accompanied with higher administrative costs, impose less risk and uncertainty on the shareholders while at the same time being politically more practicable (Frank, 2014; Shinkuma and Sugeta, 2016). This chapter addresses the general functioning of a CATS and its effects on the labor market from a theoretical perspective.
2.1 Functioning of a Cap and Trade System
Under a CATS, a limit or cap on total emissions is set during a freely chosen period of time that all companies operating within this system together may not exceed. According to the height of the cap, equivalent allowances are issued on the market, entitling companies to emit one ton of a certain GHG per allowance. Companies can then buy or sell these allowances at a market price, which regulates itself through the supply and demand available on the market (EDF, 2021). Economists argue that such market-based policy instruments like CATSs would be more effective and cost-saving than simple command-and-control approaches such as technology standards or quotas, as it would provide companies with more flexibility on how and when to reduce emissions (Fullerton, 2011; Orszag, 2010). Moreover, Orszag (2010) states that the decision of policymakers on how to allocate these allowances significantly influences the distribution of gains and losses among the economic actors. This assumption will be discussed in more detail in the further course of this work.
2.2 Distributional Effects of a Cap and Trade System
Environmental regulations, such as CATSs introduced by the state can lead to significant benefits like the improvement of air quality or related health outcomes for society like a lower mortality rate (Deschenes, 2018). However, emission abatement also entails costs, which are borne by different economic actors. The question of who wins and who loses under CATSs can only be answered by considering the distributional effects connected to them. They shed light on the redistribution of final gains and costs for individuals, firms and the state that go hand in hand with the introduction of a certain policy. Once the distributive effects are identified, they can be examined for their specific impact on the labor market. To this end, the three literature reviews are conducted with focus on specific variables relevant to the labor market, summarizing the effects on individual economic actors.
The study of Fullerton (2011) is the first that simultaneously illustrates all occurring distributional effects of an introduced CATS by applying a computable general equilibrium (CGE) model. The effects are shown in a partial equilibrium diagram of one single market, assuming emissions per unit of outcome to be fixed for simplicity. The diagram of Fullerton (2011) is depicted in Figure 1. The supply and demand for a good are reflected by the private marginal cost curve (PMC) of firms and the private marginal benefit (PMB) curve of consumers. The private market produces up to the point Q[0], where supply is equal to demand. However, every production of one unit of good causes a negative externality E (one unit of emission) which are not borne by the emitting company. The optimal output, taking the externality into account, would therefore shrink to Q', where social marginal benefit (SMB) is equal to social marginal cost (SMC, consisting of PMC + marginal external cost (MEC)). According to Fullerton (2011), an optimal CATS would now set the price for one allowance, corresponding to one unit of emission, such that output is restricted to Q'. The new gross price then accounts to Pg and the difference Pg - Pn to the price for allowances. As a competitive industry is assumed, firms can cover the remaining production costs, e.g., for labor, material and capital at a net price Pn. At this point pure profits are zero. From here, it becomes already apparent that the winners and losers of the introduction of a CATS are essentially determined by who bears the cost of the allowance price, thus who takes over the difference between gross and net price. Fullerton (2011) discusses the following six occurring distributional effects, the extent of which depends essentially on the cost coverage of the allowance price:
1. Loss in consumer surplus (red area A+D)
2. Loss in producer surplus (blue area B+E)
3. Cost of adjustment and transition (yellow area E+F)
4. Rise in scarcity rents (green area A+B)
5. Benefits of environmental protection (area C+D+E)
6. Effects on asset prices (indirect)
An increase in the final product price by Pg - Pn due to the introduction of a CATS can reduce consumers' surplus by the area A+D once the producers pass on the higher production costs fully to the consumers (red area in Figure 1) (Fullerton, 2011). The rectangle A thereby denotes the first order surplus reduction from the price increase, while the triangle D defines the second order reduction in consumer demand due to the price increase (Parry, 2004). The resulting burden of the price increase strongly depends on the demand elasticity in the prevailing market and the income that each consumer group spends on the relevant product.
However, not only the consumers but also the producers lose in surplus by the area B+E in Figure 1 (blue area). In the short-term industries face higher prices for the production of their goods due to the introduction of a CATS (Gray et al., 2014). As a quick reaction to the price increase for capital, output will be lowered (output effect) accompanied by receiving lower revenue (Yu & Li, 2021). At the same time, firms will shift away from capital to the other production function labor (substitution effect), where the final positive or negative effect on labor demand depends on whether the output or substitution effect overweighs (Deschenes, 2018). As can be imagined, firms that rely on a fixed capital-supply or on particular skills of their employees are less flexible in the adjustments of their production factors and therefore more constrained by the output effect (Fullerton, 2011). In the short term they will reduce their production factors capital and employment by the area E+F in Figure 1 (marked in yellow), described by Fullerton (2011) as the costs of adjustment and transition. If markets are characterized by imperfect mobility, dismissed employees bear the costs of retraining, relocation and possible longer phases of unemployment between jobs (Fullerton, 2011). However, once the transition is completed, CATSs may also increase the overall level of employment by creating jobs in green sectors, such as renewable energies or sustainable forestry. The transition from dirty to sustainable sectors and its implications for the labor market will be further discussed in chapter 4.2. In the long-term companies can pass on the costs to consumers and thus (at least partially) offset the negative effects of the increase in production costs (Fullerton, 2011).
In terms of economic benefits, the restriction of the polluting good creates scarcity rents by the area A+B, marked by the green square in Figure 1. Here, the allocation method of permits chosen by the government or any regulatory agency is crucial in determining the winners of these rents (Fullerton, 2011). If the allowances are handed out directly to the firms for free, then the total area becomes profits to the firms. If instead the allowances are auctioned, the area becomes the total revenue of the government. It could transfer money back to consumers which corresponds to a gain in consumer wealth at the expense of producer wealth (Parry, 2004). These effects are considered in more detail in chapter 4.2 and 4.3 on CATS consequences for HHs and the state respectively.
Moreover, with area C+D+E, corresponding to the sum of all MECs, Fullerton (2011) designates the gain from environmental protection. Since this effect is varying across countries, and even across people, and a measurement in gains and losses is based on subjective assessment, it will not be considered in the further course of this work. The same applies to the effects on asset prices, such as stock or land prices (Fullerton, 2011). Their value changes due to the subjectively anticipated gains and losses for consumers, firms, and the environment. Determining this effect is beyond the scope of this work.
All these effects identified by Fullerton (2011) have not been coherently theoretically or empirically studied to date. However, several studies such as Rausch et al. (2010 a, b) or Fullerton and Heutel (2007, 2010) use analytical CGE models to evaluate the cost effect of an introduction of a CATS on consumers and producers, thereby considering Fullerton's (2011) first and second distributional effect in particular.
Fullerton and Heutel (2007, 2010) were the first to assess the impact of a carbon tax1 on both the uses and sources side of income by showing the change in commodity and factor prices connected to an exogenous tax increase. According to the authors, a carbon tax could increase the final product price and thus has an effect on the purchasing power or well-being of the consumers (uses side). At the same time, producers may be affected through higher factor prices, especially playing a burden on capital-intensive industries (sources side). Since the model of Fullerton and Heutel (2007, 2010), similar to other CGE models, assumes factor inputs of labor and capital to be fixed in firms' production function, it does not provide a realistic assessment of economic consequences of a climate policy. However, it demonstrates the expected effect on output and factor prices and thus is helpful in determining the burden borne by different economic actors in the further course of this work (Fullerton, 2011).
Based on the tax incidence model of Harberger (1962), Fullerton and Heutel (2007, 2010) derive an analytic CGE model by assuming a competitive two-sector economy that produces two different goods. While the clean sector X uses capital Kx and labor Lx to produce its final good, the dirty sector Y additionally produces with pollution, named as input Z. Considering the EU ETS, the dirty sector would be the sector which is covered by the scheme. This results in the following production functions for both sectors X and Y with constant returns to scale and in the resource constraints K and L with fixed amounts of capital and labor:
X = X (KX, LX) Y = Y (Ky, Ly, Z) (2.1)
K = Kx + Ky I = Lx + Ly (2.2)
To derive equations of change, equations 2.1 are log-linearized and differentiated to receive (Fullerton & Heutel, 2007, 2010; Fullerton & Metcalf, 2002):
Kx^kx + KyKky = 0 L xXlx + L yXly = 0 (2.3)
Variables marked with a hat signal a proportional change (e.g., Ky = dKy/Ky) and Ay denotes the fraction of factor i (capital or labor) which is applied in the production of sector j (dirty or clean). The clean sector X only pays the wage price w for labor and the rental price r for capital, facing no further taxes on factor inputs as well as no price for pollution. Thus, it changes its factor demand according to a change in factor prices, the extent of which is defined by the elasticity of substitution in production Ox. By differentiating and rearranging oX the authors receive: (Fullerton & Heutel, 2007, 2010):
Kx - L x = Ox (pl - ) (2.4) where p[ = w and = r by assuming zero taxes on both factor inputs. Thus equation 2.4 can be rewritten as:
Kx - L x = Ox (w - f)
The dirty sector Y additionally pays the price for polluting px = Tz, which is equal to a tax on emissions in Fullerton and Heutel (2007, 2010) but can also be equated with a price per allowance in a CATS. Its factor demand choices are mainly influenced by the Hicks-Allen elasticities of substitution eij between the two factors i and j, which account for “the percentage change in the ratio of inputs due to a 1 % change in the ratio of their prices” (Thompson, 1997). This elasticity has a positive value once two inputs are substitutes and a negative value when they behave as complements (Fullerton & Heutel, 2007, 2010).
Following the calculation of choices among three inputs in Mieszkowski (1972), Fullerton and Heutel (2007, 2010) then denote 0yk = the revenue share paid to capital in sector Y and similarly define 0yl, 0yz, 0xk, and 0xl for the other factor inputs and sector X. The authors further assume that each sector fully splits its revenue among the input factors, so that 0xk + 0xl = 1 and 0YK + 0YL + 0YZ = 1. In a next step, these three input demand functions of sector Y are fully differentiated and divided by the appropriate input level, yielding equations for Kyt y and Z. Then, subtracting one of the input demand functions from the two others and inserting the Allen elasticities, the authors obtain:
Ky - Z = 0yk (eKK - ezK) r + 0yl (eKL - ezL) w + 0yz (exz - ezz) 7z (2.6)
L y - Z = 0yk (eLK - ezK) r + 0yl (eLL - ezL) w + 0yz (eLz - ezz) tz (2.7)
Through the assumption of perfect competition (price of each input is equal to its marginal revenue) and constant returns to scale in production (pxx = rKx + wLx; pyY = rKy + wLy + Tzz) a total differentiation of each sector's production function and rearrangement of the equations lead to (Fullerton & Heutel, 2007, 2010):
X = 0 xkKx + 0xlL x (2.8)
Y = 0ykXy + 0ylL y + 0yzZ (2.9)
Finally, consumers' demand response for the dirty and clean good to a change in their prices (Py, px) is derived by inserting and differentiating Ou, the elasticity of substitution in consumption between sector x and Y (Fullerton & Heutel, 2007, 2010):
X - Y = Ou (?y - px)
For a solution to the model with equation 2.1 - 2.10 consisting of eleven unknowns, Fullerton and Heutel (2007, 2010) set good X as numeraire with px = 0. They were then able to solve the model by deriving the change in factor prices w and r and output price pY given an exogenous change in the price for allowances/change in the pollution tax Tz:
The last term in the brackets of the three equations (yk - yl) Qu, marked in red, denotes the output effect. Fullerton and Heutel (2007, 2010) conclude that a higher allowance price in a CATS will increase production costs for the dirty sector, thus increases product prices and lowers output depending on consumer demand measured by Qu. Assuming that yK > yL, whereby yk = 7^ and yl= 7^, meaning the dirty sector is capital-intensive, and the denominator
D > 0, a higher allowance tax rate Tz will lead to a decrease in return to capital r relative to the wage w (Fullerton & Heutel 2007, 2010).
The first two terms in the brackets of the three equations, marked in green, indicate the substitution effect. In the medium and longer term a higher allowance price induces the dirty sector to remain at its initial output level by adjusting its demand for labor, capital and emission input (Fullerton & Heutel, 2007). Ho et al. (2008) state that firms could also substitute sustainable capital for the more carbon-intensive capital, through investments in low-carbon technologies. This would result in labor input remaining unaffected by a price change. The extent of adjustment is defined by the values of the Allen elasticities eij. Once the output effect can be neglected through equal factor intensities (YK = YL), labor becomes a better substitute for pollution than capital as soon as eLz > eKz (Fullerton & Heutel, 2010). It is then possible for firms to offset the higher production costs incurred by increasing the production factor labor and lowering the factor capital (Yu & Li, 2021).
Finally, the equation for the commodity price change py contains of a last term 0Yzfz (marked in blue in 2.8), denoting the direct effect from an increased allowance price. Thus, the costs of pollution are passed on to the consumers' commodity price in proportion to the share of revenues paid to pollution (Fullerton & Heutel, 2010).
To summarize, the model of Fullerton and Heutel (2007, 2010) illustrates that the introduction of a price on emissions has two different effects on consumers and producers: the substitution and output effect need to be contrasted. On the one hand, production becomes costlier for firms, forcing them to raise their output prices and second, carbon pricing influences the return to factors of production (namely capital and labor). The extent of these effects is determined by the magnitude of both effects, mainly characterized through the amount of the Allen elasticities, derived in the model of Fullerton and Heutel (2007, 2010).
3 Practical Example: Introduction to the EU ETS
The EU ETS is the central policy instrument of the EU to reduce the emissions of manmade greenhouse gases, considered as the main driver of global warming and climate change (European Union, 2016). It is by today the world's largest CATS, operating in all 27 EU countries plus Iceland, Liechtenstein and Norway and currently covers more than 10,000 installations, which account for around 40 % of all GHG emissions within the EU (European Commission, n.d., a).
The legislative framework for implementing the EU ETS was stipulated in 2003 with the adoption of directive 2003/87/EC. Its scope and specifications were revised and amended several times in the following years for the continuation of the system (European Commission, n.d., a). Among other things, this directive sets the criteria for national allowance allocations and specifies how the emissions are reported and monitored (Directive 2003/87/EC).
The system began its operation in 2005. It works on the same 'cap and trade' principle as theoretically described in chapter 2.2. It sets an annually cap to the overall volume of greenhouse gas (GHG) emissions that can be emitted by installations of power plants and other energy-intensive sectors whose capacity exceeds certain levels of thermal usage within the EU. The affected firms and operators can obtain and trade emission allowances on the market, resulting in a constantly adjusting market price for emissions. With each acquired allowance the holder can emit one ton of one of the three common GHGs covered by the EU ETS: carbon dioxide (CO2), nitrous oxide (N2O) or perfluorocarbons (PFCs). The cap is reduced over time to gradually decrease total emissions according to the current EU's climate targets. These were recently tightened in 2020 with the European Green Deal, aiming to achieve a climate neutrality continent by 2050. If, at the end of the year, an installation acquired too less allowances to cover its emissions fully, it is charged a significant penalty of currently €100/tCO2, expected to increase in future. (European Union, 2016; European Commission, n.d. a)
The EU ETS has been divided into different multi-year trading periods, also referred to as phase one to four. Table 1 provides an extended overview of the different phases with its associated time period, scope and further characteristics. In each phase, the scheme has been extended and adapted at different levels. For example, while the cap was steadily lowered, the sectors, emission gases and countries covered by the scheme were expanded and the allocation method of allowances was adjusted in each phase (European Union, 2015). Therefore, a simple comparison of different estimation results for each phase without considering their adjustments would lead to distorted results.
The first two phases were mainly used to set up price setting in the carbon market and develop an infrastructure required for emissions trading (European Union, 2015). Member states received allowances in proportion to their domestic emissions which are covered by the ETS and were in turn free to decide on the total quantity of allowances issued and the allocation method used (WWF, 2021). However, directive 2003/87/EC stipulated that 95 % and 90 % of all allowances in phase 1 and phase 2, respectively, should be issued free of charge to the operators based on their historic emissions while the remaining 5 % and 10 % shall be auctioned. The free distribution of allowances, known as grandfathering, has been highly criticized by environmental economists as it gives the most free allowances to intensive emitters, which are then even rewarded for their inefficient installations (zhang et al., 2017; Ellermann & Buchner, 2007). Furthermore, Venmans (2012) emphasizes that only few states made use of the possibility of auctioning: In the end, only 0.2 % and 3.1 % of all European allowances were being sold in phase 1 and phase 2, while the rest was distributed for free (Venmans, 2012).
The strong freedom of the member states to determine the number of emission permits, which were also largely issued free of charge, led to a large over-allocation of allowances on the market in the first two phases of the EU ETS. Due to this large surplus of emission allowances, phase 1 and phase 2 were characterized by a high price volatility, shown in Figure 2. While in April 2006, the price per ton CO2 reached a peak level of €30, it dropped to under 10€/ton in the following months after the European Commission announced that total emissions in 2005 were 4 % lower than the amount of allowances distributed (Ellerman & Buchner, 2008). At the end of the first phase in 2007, the price per allowance collapsed to almost zero since companies were not allowed to take over the available allowances on the market into the second phase. This procedure, called banking, was first introduced towards the end of phase
2. However, the second phase was significantly shaped by the economic crisis, which led to unforeseeable emissions reductions and thus to a sharp drop in demand for allowances (Borghesi & Montini, 2016; European Commission n.d., b). These reasons led to the fact that at the beginning of 2013, a surplus of two billion unused allowances circulated on the market (FSR, 2021).
In order to tackle this inefficiency, the system underwent a significant reform in the third phase, lasting from 2013 to 2020, with the main aim to substantially harmonize the scheme across the EU countries. All fundamental decisions have now been taken at EU level and a uniform EU-wide cap has been introduced, which has decreased linearly by 1.74 % each year (Directive 2009/29/EC). Additionally, the auctioning instead of free allocation moved into focus with 57 % of all allowances being auctioned among the affected operators. For example, from 2013 onwards, the power generation sector was required to purchase all of its allowances in auctions only (Directive 2009/29/EC). The remaining 43 % of all issued allowances were allocated freely via a mixture of the so-called benchmarking approach and grandfathering. The benchmarking approach was mostly applied to industrial and heating sectors. It determines the total number of free allowances that each installation should receive by multiplying a firm's production quantity with the benchmark value for that product it produces (European Commission, n.d. d). The benchmark value is the average emitted tons of CO2 per product produced by the 10 % most GHG efficient installations in the years 2007 and 2008 (European Union, 2015). While in 2013, 80 % of the determined quantity via benchmarking was freely allocated to the installations in the heating and industrial sectors, the amount decreased to 30 % in 2020 and targets a value of 0 % in 2027 (European Union, 2015). With the benchmark approach, installations that produce highly efficient are rewarded by receiving all or almost all of their allowances needed for their production for free, while inefficient installations need to become active to either buy the remaining allowances or reduce emissions (European Union, 2015). Given these changes in the allocation method, in 2013, the EU has published a list of sectors exposed to a high risk of carbon leakage, which should therefore continue to receive their allowances 100 % free of charge via grandfathering (European Commission, n.d. d). Carbon leakage defines the risk of firms moving their production to other countries which are not affected by the EU ETS. A sector is exposed to a high risk if the EU ETS would increase its production costs by at least 5 % and its trade intensity with non-EU countries by more than 10 % (European Commission, n.d. d). A second and third list was applied from 2015 - 2019 and 2021 - 2030 respectively, while steadily reducing the sectors being covered by the list (European Commission, n.d., d).
To achieve the goal of climate neutrality by 2050, EU leaders had to agree on new binding EU climate targets in phase 4, aiming for a 55 % instead of the previous set 40 % reduction in GHG emissions by 2030 compared to the levels of 1990 (European Council, 2021). Therefore, phase 4 is characterized by an even stronger annual cap reduction than in phase 3, namely by 2.2 % from 2021. Furthermore, a modernization and innovation fund were set up to support the development of low-carbon technologies, especially in lower-income EU member states (European Commission, n.d., e). Both funds are equipped with the revenues from auctioning allowances in phase 4 (European Commission, n.d., e).
To tackle the high supply-demand imbalance and price volatility of allowances occurred in phase 1 and 2, as a short-term instrument the Commission postponed the auctioning of 900 million allowances from 2014 - 2016 until 2019 - 2020 (FSR, 2021). This approach is known as backloading of auction volumes. Shortly before the beginning of phase 4, a long-term solution in form of a market stability reserve (MSR) was implemented in 2019. Every time, the Total Number of Allowances in Circulation (TNAC) exceeds the mark of 833 million allowances, 24 % (2019 - 2023) or 12 % (after 2023) of the total surplus of permits are transferred to the reserve (intake rate) (FSR, 2021). In contrast, if the TNAC falls below the value of 400 million, 100 million allowances will be transferred back from the reserve to the market via auctions (FSR, 2021). In addition, from 2023 on, all allowances in the reserve that exceed the auction volume of the previous year will be invalidated and permanently annulled (Johansson, 2021). The success of these measures is shown in Figure 3. With the beginning of back-loading in 2014, for the first time the number of allowances entering the market corresponded to the actual quantity of emissions. However, the TNAC still exceeded the number of verified emissions, owed to the longstanding over-supply of allowances. Since the introduction of the MSR in 2019, the TNAC has fallen sharply as a result of moving allowances into the reserve (Johansson, 2021). The MSR is expected to stay below the threshold of 833 million issued allowances (and thus below the threshold where it gets active) from 2023 on. Johansson (2021) claims that through the MSR the supply of permits becomes endogenous. Underscoring the findings of Perino (2018) and Silbye and S0rensen (2019) he states that this reform can be described 'to puncture the waterbed', meaning that a country's reductions in emissions finally cause total emissions within the EU ETS to decrease instead of simply redistributing emissions within the EU. However, Figure 3 also shows that from 2026, the projected emissions of all covered installations will exceed the cap for the first time, forcing companies to either accept higher penalties or stronger invest in emission reduction procedures.
The reform of the EU ETS in phase 3 and 4 is also reflected in a more stable allowance price development, visible in Figure 4. With the start of phase 4 and the accompanying of stronger emissions reduction targets, the price rose from €32 in mid-December 2020 to €54 in mid-June 2021 within half a year. This sharp increase is a harbinger of the strong price development that can be expected in the future due to the artificially induced shortage of allowances by the MSR. In summary, the EU ETS has evolved from an initially inefficient policy instrument to a forerunner model for other countries. Especially its great market size of currently 30 involved countries and its rapid adaptability within the trading phases distinguishes it from other successors. However, increasingly stringent EU emission reduction requirements, as they were introduced specifically in the fourth phase, pose strong challenges to the companies, caused by an annually lower cap set and an associated sharp rise in the price of allowances. In order not to exceed the cap, firms are left with two main options: either reduce their emissions by investing in energy-efficient production processes or reduce their production costs in other ways, e.g., by cutting wages or dismissing employees. Their decision ultimately affects not only their profits but also the wealth of employees and the state.
The next chapter presents and critically examines the current literature that empirically and theoretically investigates the impacts of a higher CO2 price in the EU ETS on the wealth of state, HHs and firms.
4 Literature Review
Since the EU ETS is a complex pattern of 31 countries involved, an evaluation in form of a literature review seems the most appropriate method to gain a broad perception of the economic costs borne by the three economic actors and their effect on the labor market. Thereby, 51 studies differing in their paradigms and methodologies were analyzed and their results contrasted. Individual studies are presented in more detail at the author's discretion that contribute to the argumentation of the effect under consideration.
To the author's best knowledge, by date there is no literature review that focuses exclusively on the labor market effects of the EU ETS and its consequences for all three economic actors, firms, HHs, and the state, separately. One possible explanation could be that labor market effects is a very broad term under which various individual effects and thus output variables can be observed. Moreover, since the EU ETS is the largest CATS in the world, cross-country labor market effects are difficult to measure as they depend on various country-specific factors. This is why previous literature reviews such as Joltreau and Sommerfeld (2018) or Verde (2020) rather focus on empirical studies for one single actor (in their case: firms). To evaluate the impacts of the distributional effects on the labor market elaborated by Fullerton (2011) and presented in chapter 2.2, the analysis will focus on literature that assesses the following variables, either on average EU-level or single country level:
1. For firms : change in number of employees and profitability (measure of competitiveness)
2. For HHs : effect on income and change in wages (measure of poverty and inequality)
3. For state : change in GDP growth (measure of economic growth)
The analysis begins with the firm side, as they are the emitters directly affected by the introduction of the EU ETS and the associated purchase of emission allowances.
4.1 Effects on Firms
The EU ETS currently covers all installations of power plants and other energy-intensive sectors whose capacity exceeds certain levels of thermal usage within the EU. Once the EU ETS was implemented in 2005, it immediately imposed constraints on firms' production whose effects depend on various factors such as firms' factor composition in production, their market position in international competition or consumers' demand response. In the following, the impacts of the EU ETS on the competitiveness of companies are examined and discussed. Thereby, two different competing economic theories are used to evaluate the results.
4.1.1 Context
A common procedure to assess the labor market effect of environmental regulations such as the introduction of the EU ETS on firms is the evaluation of changes in their competitiveness (Joltreau & Sommerfeld, 2018). However, the term is not uniformly defined - while Bristow (2005) refers to competitiveness as a firm's or sector's potential to withstand competition in a market, to grow and be productive, Berger (2008) understands competitiveness as a firm's ability to increase market share or profits. Common to all views is that competitiveness enables companies to increase their market influence rather than being displaced by other competitors. The introduction of a CATS, which Fullerton (2011) found to be associated with various distributive effects, has also significant implications for firms' competitiveness. In particular, the loss in producer surplus and the rise in scarcity rents through grandfathering permits are found to have far-reaching effects on firms' employment and profitability. These two measures significantly determine a company's market influence (Dechezleprêtre & Sato, 2017). Table 2 illustrates the three downstream competitive effects that occur for regulated firms under the EU ETS, elaborated by Dechezleprêtre and Sato (2017). According to the authors, the firstorder effect occurring after the introduction of an ETS, is a considerable increase in companies' production costs depending on the price and the allocation method of allowances implemented. The authors differentiate between direct cost effects through increasing material costs and indirect cost effects through rising electricity and other input costs necessary for production. Without any counteracting activity of the firm, this first-order effect refers to the loss in producer surplus, identified in chapter 2.2, since firms' production costs increase while the product price at which they are remunerated remains fixed in the short-run. In the medium- and long-run this cost impact induces different responses of the firms as part of the second- order effect. For example, firms may answer by an adjustment in production volume, an increase in product prices (by the term 0yztz shown in equation 2.8 in chapter 2.2), a substitution away from the more costly production factor capital to labor, or by a stronger investment in abatement (Dechezleprêtre & Sato, 2017). These emission abating activities will then in turn influence various outcomes in the economic, technological, international or environmental field, defined as third-order effects and listed in detail in Table 2 .The following literature review will mainly focus on the economic outcomes under firms' regulation in the EU ETS, namely the change in the number of employees2 (equal to firms' employment demand) and in firms' profitability (through direct and indirect measures such as profit margin and turnover).
According to Ton (2009), the quality of a product is essentially dependent on the input factor labor, which in turn strongly influences the profitability of a company. If the number of employees decreases due to increased production costs, the profitability of the company may also decrease, which may lead to further layoffs or even the permanent closure of the company (Ton, 2009). Therefore, the two variables are considered interdependent in the further course of this literature review.
4.1.2 Methods and Data Samples in the Literature
Many studies were found examining theoretical ex-ante and empirical ex-post impacts of international trading schemes on firms' competitiveness for the U.S. (e.g., Rausch et al., 2010 a) or for China (e.g., Yu and Li, 2021). Table 3 presents a review of the literature that focuses specifically on the changes in firms' employment and profits related to the introduction of the EU ETS. For the purpose of this literature review, empirical ex-post studies were selected for closer examination from the wide range of research, as they are considered to stronger rely on real-time information rather than unproven assumptions and guesswork (Samset & Christensen, 2015). For ease of reference, the studies were subdivided according to their regional and sectoral sample coverage and arranged in ascending order by year of publication.
When studying Table 3, it is striking that only a minority of studies employ literature reviews (Verde, 2020; Joltreau & Sommerfeld, 2018), Two-Stage Least Squares (2SLS) models (Anger & Oberndorfer, 2008) or OLS Panel-Data regressions (Commins et al., 2011). The econometric Difference-in-Differences (DID) method emerges as the main and most suitable method to empirically evaluate the performance of the EU ETS. Therefore, this approach is now explained in more detail.
Econometric DID approaches compare average causal effects of a specific intervention or event, such as the introduction of the EU ETS, on an outcome variable, e.g., number of employees, for regulated and non-regulated firms, defined as treatment and control group, respectively (Verde, 2020). It thereby studies outcomes for both groups over a time interval striding at least one period before and one after the implementation of the EU ETS. The average treatment effect (aArr) of the firms being regulated under the EU ETS can then be calculated by:
aATT = E [Yit (1) - Yit (0) | ETSi = 1] (4.1)
where ETSi = 1 refers to plant i being regulated under the EU ETS and thus belonging to the treatment group, while ETSi = 0 refers to non-regulated firms in the control group respectively (Jaraité & Di Maria, 2016). Yit (t) and Yit (t) denote the outcome variable for plant i and time t conditional on being implemented under the EU ETS (t=1) or not (t=0) (Löschel et al. 2018). One major problem occurring with causal interference, is that the second part of the equation, E [Yit (0) | ETSi = 1], is not observable, meaning that the outcome of ETS firms which have not participated in the scheme is unknown (Jaraité & Di Maria, 2016). To solve this limitation, the DID approach makes use of the observed outcomes of the unregulated control group. Therefore, it first calculates the selection bias, which denotes the difference in outcomes between the treatment and control group before the implementation of the EU ETS. This bias is subtracted from the calculated difference between both groups after implementation to receive the average treatment effect aATT on the treatment group (Verde, 2020; Jaraité & Di Maria, 2016).
However, for the DID method to provide unbiased estimates the following key assumptions must be fulfilled: First, the EU ETS must only affect regulated companies (Stable Unit Treatment Value Assumption - SUTVA), thus eliminating spillover and general equilibrium effects for unregulated firms (Wagner et al., 2014). Second, in absence of the EU ETS, the trends in the variable of interest for the two groups would develop in parallel (Common Trend Assumption) (Verde, 2020; Angrist & Pischke, 2008). However, both assumptions are only fulfilled to a limited extent for the EU ETS: Regulated and non-regulated firms differ in many factors like their size or production capacity, which would significantly disturb the parallel trend development without the EU ETS (Stuart et al., 2015). Moreover, the introduction of the EU ETS has also affected non-regulated firms, e.g., through increased indirect production costs in form of higher electricity prices (Dechezleprêtre & Sato, 2017).
To relax the assumptions on their estimates and reduce the potential bias, most DID approaches use semi-parametric matching estimators to construct similar counterfactuals to their observed regulated firms under the EU ETS (e.g. Wagner et al., 2014; Jaraité & Di Maria, 2016 or Löschel et al., 2018) Most studies in Table 3 are found to apply a nearest neighbor (NN) propensity score matching.3 Propensity score defines the estimated probability for a firm being regulated under the ETS as a function of pre-treatment observable characteristics X, P (ETS = 1| X) (Stuart et al., 2015). Under NN matching a regulated firm is matched with a counterfactual from the control group that is closest in terms of the propensity score (Caliendo & Kopeinig, 2008). Unlike Colmer et al. (2020) who uses one-to-one NN matching, Jaraité and Di Maria (2016) and Löschel et al. (2018) match each firm in the control group more than once with a firm in the treatment group. According to Colmer et al. (2020) matching with replacement increases the average quality of matching and decreases the possible bias in the asymptotic distribution of the matching estimator, which is why he strongly advocates the use of this method in subsequent studies.
Following Heckman et al. (1997, 1998) all studies using the DID approach with NN matching summarized in Table 3 estimate the following generalized matching DID estimator:
aATT = E [Yi1 (1) - Yi0 (1) | ETSi = 1] = ±zj&i {(y#i(i) - YJto(oj) - £fce/o Wjk (ykti(o) - ykto(o))}
(4.2) where I1 and I0 define the set of regulated and matched non-regulated firms respectively, while j is the index for regulated and k the index for non-regulated firms (Wagner et al., 2014). N denotes the total number of firms that the treatment group consists of. The size of the weight wjk depends on the matching technique used. It is placed on firms in the control group to create similar counterfactuals to regulated firms (Wagner et al., 2014; Jaraité & Di Maria, 2016).
In addition to the frequently applied DID method, a recurring pattern in the data sample used is striking. Most listed studies use micro-level data on the firm - level or in some exceptions even plant-level such as Wagner et al. (2014) and Klemetsen et al. (2016). For reasons of data protection but also due to large data volumes, survey results are often published as aggregates, which however result in a significant information loss. Only micro-level data allow to explore detailed interactions between variables (IPUMS International, n.d.). Moreover, the considered time period of the sample is central for the analysis of results, since the regulatory requirements changed with each phase of the EU ETS (see chapter 3). It is noticeable that the current studies mainly examine the period of the first and second phases. Only a few also consider data from the beginnings of the third phase (De Bryn et al., 2015; Klemetsen et al., 2016). The impact of these observation periods on estimation results is considered in the next chapter.
4.1.3 Evidence: Impact of EU ETS on Employment and Profitability
The overview of the empirical literature in Table 3 already shows substantial differences in their sectoral and regional scope as well as in their output variables of interest considered. A large part of the studies focuses on the measurement of cross-country, cross-sector effects. Linking the AMADEUS database, a comprehensive dataset of more than 21 million European companies, Commins et al. (2011), Abrell et al. (2011) and Chan et al. (2013) measure the impact of the EU ETS on profitability and employment across firms in all EU member states. The study of Commins et al. (2011) solely focus on the first trading phase of the EU ETS from 2005 - 2007. Across EU firms, they find a significant negative effect on productivity and profits (-3.2 % and -4.7 %) but a significant positive effect on employment (+1.5 %). While Abrell et al. (2011) find no significant impact on profit margins and only a small but significant negative effect (-0.9 %) on employment, Chan et al. (2013) find no significant effects on employment or competitiveness in European firms under the EU ETS. Both studies are very similar in terms of approach and data sources used, a significant difference however, is that Chan et al. (2013) draw on a more extensive data set and make sure to match regulated firms with a counterfactual firm from the same industry. The authors argue that this method would omit potential bias which usually arises when time-variant differences among industries are not
[...]
1 It must be noted here that a CATS including tradable emission allowances that are fully auctioned to firms is assumed to have the same distributional effect as an emission tax once both induce equal emission reduction targets (Parry & Williams, 2005).
2 For the definition of employment, most studies follow the definition of ILO, whereby a person of working age is employed if he or she is engaged in a (self)paid activity within a certain short period of time, e.g., one week (ILO, n.d.).
3 Although various matching algorithms exist, see Caliendo & Kopeinig (2008) for an overview.
-
-
-
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X. -
-
¡Carge sus propios textos! Gane dinero y un iPhone X.