European Emission Trading System: A Market-Based Solution to a Market Failure
Environmental Economics emerged as a response to market failures. In the classical economic theorem, the markets result in the most efficient outcome and the invisible hand of the market always finds the right allocation of resources. However, this theorem based on some assumptions, namely, the atomistic participants (no monopoly), perfect information, no transaction cost, no externalities, no taxes, no public good, no common property or no other distortions between costs paid by buyers and the benefits received by the sellers.[1] As we know that there has never been a perfectly competitive market. For this reason, markets need external interventions in order to fix certain problems. One of the most important market failures is the carbon emissions ensuing global warming. In order to fix this externality, a cap-and-trade system was developed within the framework of the Kyoto Protocol. Basically, it is a system limiting the Green House Gas (GHG) emissions. Companies receive or buy allowances (permits) to emit a certain amount of GHGs. If they do not use these permits (emit less than their permits), they can trade them. The European Emission Trading System (ETS) is the first example of this cap-and-trade system which, basically, sets the property rights to emit.[2]
In this paper, we shall try to look at the basic fundamentals of cap-and-trade, the command-and-control and carbon tax systems. Afterward, we will look at the European Union Emission Trading System (EU ETS) and cover the failure of this system i.e. the price volatility. We shall try to understand the reasons behind the price volatility. Lastly, we will look at the impacts of EU ETS on the gross output, employment, and exports of the covered sectors.
What is a cap-and-trade system? What is its difference from other systems which also intend to reduce GHG emission e.g. command-and-control and carbon tax?
The cap-and-trade system is a market-based approach to fix a market failure, namely, emissions. In order to find the social optimum, the social cost of the GHG emission (in this case the global warming) should be internalized in the cost structure of the firm.
In a cap-and-trade system, this externality is internalized by setting an upper limit (a cap) on aggregate emission of specified plants, and then by trading the allowances among the firms. “Such trading is conducted through the sale and purchase of allowances, which are issued in an amount equal to the aggregate cap. Allowances can be acquired either through free allocation or by purchasing through auctions or from others through trading.”[3]
This goal could also be achieved through taxes (by setting the tax at the cost of emissions) or through non-market solutions such as command-and-control. In the latter case, the regulatory body decides the maximum amount of emission that each firm is allowed to. Yet, among all these three systems, the command-and-control is the farthest to ensure the economic optimum because it does not take into account the different abatement costs of different firms. The carbon taxes fix the price of the emission and let the firms adjust the quantity of their abatements; whereas, in the cap-and-trade system, the regulatory body release the permits, therefore, it fixes the quantity abated and let firms determine the permit price. The permits would be allocated to the firms for free (which is also known as grandfathering) or through auctioning where the marginal abatement cost of the last bidder sets the price. [4]
How the prices are set in the permit market: the theoretical foundations
The theoretical foundations of prices in the permit market are based on the basic principles of supply-demand interaction. As we discussed above, the government sets the quantity of emissions. If the combined emission is lower than the cap (which also means combined abatement is larger than what government intended), not all emission permits are sold, the prices are too high and there is an excess supply of permits. Then some firms will be willing to lower the prices to sell them. As a result, the demand for permits will increase, and the optimum price and quantity of permits would be set. Likewise, if the combined emission is larger than the cap (which also means the combined abatement is lower than what is intended), there are fewer permits available than the amount of pollution. In this case, some firms don’t have a permit for their emissions which results in excess demand for permits – a shortage. Consequently, some firms will be willing to increase the prices, the quantity demanded will decrease and the market will again reach the optimum price and quantity. In both cases, the marginal abatement cost of all firms (that is the cost of the last unit of carbon that is not abated) will eventually be equal to the optimum permit (allowance) price. The permit system requires the knowledge of industry marginal abatement cost curve and there is a possibility of learning it from observing permit prices.[5] However, as we will see in the case of the European Emission Trading System, the marginal abatement cost (MAC) theorem is not sufficient to understand the cap-and-trade system. In many cases, MAC explains only a small portion of prices. The EU Allowance (EUA) prices have spiked up and down many times and the MAC theorem was not sufficient to understand these ups and downs.
Moreover, there are certain challenges for the implementation of the permits program which would be alternative explanations for the price volatility. First of all, there are enforcement issues. It is easier to enforce the permit system when there is a limited number of firms. Besides, monitoring would be costly. Secondly, the liquidity of the market matters. If there are a few power plants (or firms), they can have market power and withhold the permit to increase the prices or refuse to sell permits in order to block the new entries to the market. And lastly, there is a problem with non-uniformly pollutants. In the case of some pollutants like CO2, the cumulative and global emission matters. However, in the case of some pollutants like SO2 or NOx, where the pollutants are emitted matters. For these type of pollutants, different regulations are required.[6]
European Emission Trading System (ETS)
Based on the cap-and-trade mechanism explained above and following to European Union’s commitment to the Kyoto Protocol, EU Emission Trading System (ETS) was set up in 2005 and so far, has implemented in three phases. Phase I was the trial phase from 2005 to 2007. Phase II was from 2007 to 2012 and Phase III will be in force until 2020. ETS is the world's first and biggest international emissions trading system making up over three-quarters of international carbon trading. It covers the trading of carbon dioxide (CO2) from power and heat generation, energy-intensive industry sectors including oil refineries, steelworks, and production of iron, aluminum, metals, cement, lime, glass, ceramics, pulp, paper, cardboard, acids, and bulk organic chemicals; nitrous oxide (N2O) from production of nitric, adipic and glyoxylic acids and glyoxal and perfluorocarbons (PFCS) from aluminum production. Yet, small industries, some sector plants below the certain size and aviation sector (until 2023) are exempted.[7]
Allowance Price and Price Volatility in ETS
As we mentioned above, initially the allowances have been either distributed to the treated firms for free or auctioned. During the trial period of ETS (Phase I), most of the allowances were allocated for free; whereas starting from Phase II, some of the allowances (permits) have been auctioned. However, during Phase III, the permits will be mostly auctioned. According to Ellerman et all[8], free allocation of allowances[9] paved way firstly for windfall profits and secondly for competitive distortions (e.g. market power). During Phase II, the free allocation was terminated for the electric utility sector excluding some coal-dependent Eastern European countries; and a more gradual termination has been designed for non-electric industrial sectors due to the concerns of international competition.
As we can see from the graph above, there were two separate permits (allowance) markets. The Phase I allowances were only permitted to be traded during Phase I and could not be carried to Phase II. However, for Phase II and III, with the introduction of the banking system, one single allowance market was created. According to Ellerman et all. (p.99)[10], the banking system’s role was strengthened in order to establish a floor price. The prices hit the zero lower bound during Phase I. In order to overcome this problem, during phase II the surplus allowances were banked and permitted to be carried to Phase III.
Graph 1: The graph shows the price volatility of the ETS system[11]
During both the first and second phases, price volatility was a big concern. In 2005, the allowances for Phase I was around €5 and €10. Then they increased up to €35 and stayed in this range until May 2006. During summer 2006, the Phase II allowance prices dropped by fifty percent to €15; whereas Phase I prices hit the zero lower bound. Just before the 2008 financial crisis, the prices were at €30 range. With the crisis, it decreased again by fifty percent. After a brief recovery, it continued to fall, and it fell to €7-8 range in 2012. In 2013 at the beginning of Phase III, the allowance prices were around €3.65. The lower prices indicate that the European emission was below the cap. There is not so much demand for emissions and the abatements are higher than expected. Ellerman et all[12] asserts that there are two different views regarding the low prices. One view claim that much-lower-than-expected prices are the indicators of flaws of ETS. The other view says that the low prices show that the ETS is working proving that the abatement cost is cheaper than expected. Let’s elaborate on the possible reasons for price volatility and how much of the price volatility can be explained by the market fundamentals i.e. the MAC theorem.
The Reasons behind the Price Volatility: Market vs. Out-of-Market Dynamics
Hintermann[13] searches for the reasons for the high prices during the first half of Phase I. He is trying to find out that how much the basic principles of the economics i.e. the MAC can explain the price volatility. His research focuses on the MAC function of the power producers. He sets up an economic model that specifies the allowance price changes as a function of a set of widely accepted price drivers (fuel prices, temperature, reservoir levels, economic indicators and announcements of verified emissions), under the assumption of efficient markets.[14] By this equation, he is trying to find out if the price changes are the result of the fundamentals of the MAC. He is specifically searching for the reasons for high prices during the first half of Phase I and Phase II. Therefore, he is separating the pre-crash and post-crash periods from each other i.e. the period before and after high prices.
According to the table below, he finds out that the market fundamentals explain a significant portion of the variation in EU allowance (EUA) price changes in the post-crash period. However, this is not the case before the price crash except for gas prices and Nordic reservoir levels. Coefficients on fuel prices, temperature, and availability of hydroelectric power are statistically significant and have the expected sign for the period after the price adjustment. In the post-crash period, market fundamentals drive the allowance price in a nonlinear way. Although gas prices and Nordic reservoir levels contribute to the explanatory power in extension, the most important determinants of EUA price changes are lagged price changes.
Table 1: Hinterman’s Regression Analysis [15]
Yet, there is still a gap that cannot be explained by market fundamentals. He[16] says that the possible reasons of the allowance prices’ deviation from the marginal abatement costs would be the CO2 bubble, the market power of firms, firms hedging against stochastic (randomly determined) emissions and the allocation updating. Accordingly, a CO2 bubble would be the result of bullish market reports, price manipulation by dominant market players, or some traders who did not recognize the fundamentals of the market and expected higher prices in the future. Moreover, there might have been a market power such that some dominant players set the price in order to maximize their own profits. However, considering the main recipients of the allowances were the power generators, they would rather prefer to depress the market prices. Furthermore, if firms were unable to effectively control their emissions in time for Phase I, either because abatement was simply not feasible in the short run as firms were locked into long-term contracts, or because emissions are a function of stochastic (randomly determined) output. Then, firms have to hedge against the probability of having to pay a penalty which increased the prices higher. Lastly, the EU updated its allocation plan according to the 2005 verified emissions rather than the historic or projected level of emissions. Such allocation ‘‘updating’’ resulted in a disincentive to abate, and formally show that the permit price will be greater in this case than if allocations were fixed based on historic emissions. While all these four explanations are likely, it is not possible to know what exactly caused the separation of the prices from the marginal abatement cost.
Creti and at all[17] try to explain the low prices during Phase Il. According to their research, it is possible to explain the carbon prices as a function of oil price, equity price index, and the switching price between gas and coal. Through their regression analysis, they come up with the equations below:
Carbont = 4.8852 + 0.7117Brentt – 0.7676 Eurext + 5.66.10–5switcht (R2=0.3285) for Phase I
Carbont = ?0.3251 + 0.5338Brentt + 0.5009 Eurext + 0.0057switcht (R2=0.7305) for Phase II.
The Brent represents the Brent oil prices, Eurex represents the equity price index (an indicator of economic activity), and the Switch represents the switching price between gas and coal. According to their findings, the downward trend of permits price (the one that started in August 2008) is the result of the decline in oil prices and ensuing decline in gas prices. The decline in gas prices triggered the switching from coal to gas during Phase II which also resulted in a reduction in emissions. The decline in natural gas prices depressed the carbon price. However, during the period started with the first quarter of 2009, the economic crisis has further depressed the observed carbon futures market price. Although European energy demand sustained for a while, the production cutbacks in steel, cement and glass industries (due to decline in demand) caused downward pressure on carbon prices.
Graph 2: Creti et all. graph showing the gap between market fundamentals and the actual prices[18]
The model they set explains the price volatility to some extent. Their findings show that the equilibrium relationship (Eu Allowance prices equal to the market MAC) exist for both phases, with an increasing role of fundamentals in Phase II. In particular, while all the considered explanatory variables—namely, oil price, equity price index, and the switching price between gas and coal—are significant long-run determinants of the carbon price in the second phase of the EU ETS, the switching price does not play a key role in the first phase. However, from the end of 2009 onwards (as it can be seen from the graph above) the equilibrium price implied by the fundamentals stayed systematically below the observed one. They have two possible explanations for this discrepancy. Firstly, the market has probably overreacted to the uncertainty of the international climate policy. Secondly, there was a lot of fraud in 2009. The “carousel fraud” (failure by allowance sellers to pay back to the Member States the VAT they collect) and several phishing attacks lowered the exchange volume and thus the carbon price.
According to Koch et all.[19], theoretically the permit prices should reflect market fundamentals related to the marginal costs of abatement. Fuel switching in the dominant power sector is considered to be the single most important abatement method in the EU ETS. Consequently, in an efficient market, prices for input fuels are expected to determine EU allowance prices. In addition, exogenous factors such as economic activity or weather conditions are identified as relevant price fundamentals, since they determine business-as-usual emissions, i.e. the need for abatement. However, in many cases, those market fundamentals are not enough to explain the price volatility. Most of the papers’ common finding is that the identified marginal abatement cost drivers had only a limited influence on EU Allowance (EUA) price formation. Therefore, Koch et all. claim that there would be four possible explanations for price volatility apart from the market fundamentals. These are the long-lasting economic crisis, overlapping climate policies more specifically the deployment of renewable energy sources (RES), the large influx of Certified Emission Reductions (CERs) and Emission Reduction Units (ERUs) in the EU ETS during Phase II. According to their hypothesis, they come with the equation below
EUAt = β0 + β1Switcht + β2Economyt + β3Wind/Solart + β4Watert + β5CERt - 1 + εt (1)
EUAt = β0 + β1Gast + β2Coalt + β3Economyt + β4Wind/Solart + β5Watert + β6CERt-1 + εt (2) [20]
Table 2: Koch et all’s regression analysis[21]
Table 2 above, rows (1) and (2) show the results for Equation (1) based on switching price, while rows (3) and (4) show the results for Equation (2) based on the individual fuel prices. The respective rows only differ with regard to the economic activity proxy used. The different regressions reveal a number of interesting conclusions. First, for the fuel variables, they find no clear-cut evidence that the abatement costs of fuel switching are reflected in EUA prices. Although the change in the switching price shows the expected positive coefficient estimate, its statistical significance is ambiguous, depending on the economic activity proxy used (ESI vs. stock index). Second, the estimation results document that variations in expected economic development exhibit a strong influence on EUA price fluctuations. Ceteris paribus, a 1% decrease of the ESI is associated with a decrease in the EUA price of approximately 1.2%. Third, the development of intermittent renewable energy deployment helps explain EUA price movements. However, the magnitude of the price elasticity of wind/solar growth (-0.11% and -0.14%) is relatively small (specifically, in comparison to the price elasticity of economic activity). Finally, there is a statistically significant negative influence of the issued CERs. However, given the tiny coefficient estimate, the economic significance of CERs seems rather limited.
Graph 3: Decomposition of the adjusted R2 based on the estimation of Eq. (1) in a row (1) of Table 2 into contributions of the different marginal abatement cost (MAC) drivers based on Shapley values.[22]
In short, Koch et all’s selected abatement-related fundamentals explain only about 10% of the variations of EUA price changes. The figure above shows a decomposition of the R2 into contributions of the different marginal abatement cost drivers. Accordingly, approximately 40% of the explanatory power of the model can be attributed to variations in economic conditions and around 23% of the R2 relate to the growth of wind and solar deployment. In other words, economic activity and wind/solar deployment are the most important abatement- related drivers of EUA price dynamics. In contrast, the influence of fuel switching is rather small. Overall, the key finding of their analysis is that EUA price dynamics cannot be solely explained by marginal abatement cost theory.
Conclusion
To sum up, the integration of the banking system was good for establishing a floor price. The allowance prices hit the zero lower bound during Phase I mostly because the permits could not be carried to the future. However, the banking system has its own flaws too. It would disincentive the innovations and further abatements. Knowing that allowances would be carried to further periods, the firms could reallocate their emissions across time rather than investing in the cleaner or more energy-efficient technology. As Koch et all.[23] put it the depressed permit prices are not likely to provide sufficient incentives for low-carbon technological investments.
The theory of marginal abatement cost does not hold many times. The theory tells us that abatement starts where the marginal abatement cost is the smallest. And the allowance price would be eventually equal to the market marginal abatement cost. However, many of the studies show that in the EU ETS system the MAC does not necessary the only explanatory variable. The papers we discussed above tell us that the allowance prices would be written as a function of oil prices, switching prices from coal to gas (for power producers), economic activity, weather, deployment of renewable energy resources and so on. These fundamentals explain the price changes and price volatility to some extent and apparently their importance increased during Phase II. Yet, they are not sufficient to explain the whole story behind the price volatility. On the one hand, we know that the MAC theory can explain only 10% of price changes. On the other hand, we also know that (for the time being) it is not really possible to address the real causes of the price volatility either.
Lastly, because we know that the MAC theory and market fundaments importance increased during the Phase II, there might be a correlation between the penetration of the banking system and the increased importance of the market fundamentals. However, there is no research conducted to prove this hypothesis.
Last but not the least, the graph below shows the latest price changes in the EU ETS system. As we have discussed above, oil prices are one of the fundamental determinants of allowance prices. As we all know, the oil prices increased up to $80 in the last couple of months and declined to $60 level again. We might claim that the sudden increase in the allowance prices in October 2018 would be due to the Iranian sanction and ensuing oil price increase.
Graph 4: EU ETS Price Chart[24]
Please Read the Following Part as an Annex to the Paper
The Impact of ETS on Industry Employment, Output and Exports
Apart from the price volatility, one of the issues raised around the EU ETS is about its possible negative impacts on employment, gross output and exports. In the short run, the companies could not innovate or could not invest in energy efficient technology. Therefore, it is believed that they have to reduce their gross production to catch the emission cap. Yet, this criticism would be true only if the EU ETS system was a command-and-control system. Let’s elaborate this criticism based on a research about the German manufacturing firms.
A research conducted by Petrick and Wagner[25] investigates if there is a direct impact of ETS in CO2 emission and if ETS had a negative impact on employment, output and exports of the manufacturing firms in Germany.[26] They found out that the treated firms abated one-fifth of their CO2 emissions between 2007 and 2010 relative to non-treated-firms. Besides, there is no evidence to support the fear that carbon pricing (market) is reducing the employment, gross output and exports of the firms.
CO2 Emissions and CO2 intensity
According to the table 3 below, Petrick and Wagner[27] assert that while there is no significant abatement during Phase I, the treated firms reduced their CO2 emissions between 25% and 28% in comparison to non-treated firms in Phase II. Likewise, while there was no significant change in CO2 intensity during Phase I, firms reduced their carbon intensity between 18% and 30% faster than the non-treated firms in Phase II. (The findings are significant at 5% level.)
Table 3: The CO2 emissions and intensity in Phase I and II[28]
Employment, Gross Output, and Exports
In terms of employment, Patrick and Wagner’s research[29] shows that none of their findings either in Phase I or in Phase II is statistically significant. Therefore, there is no evidence to support that carbon pricing reduces the employment rate. Regarding the gross output, there is no statistically significant result for Phase I. However, during Phase II, the output of the firms, contrary to general belief, increased between 4% and 7%. Lastly, their study enabled them to reject the hypothesis that the EU ETS caused regulated firms to reduce their overall exports. For further details please see table 4 below.
Table 4: EU ETS’ impact on employment, gross output and export[30]
The most interesting part of this research is its findings of how the firms reduced their CO2 emissions e.g. Is it by increasing their energy efficiency or decreasing their electricity consumption? Accordingly (p.28), the treated firms decreased their natural gas and oil consumptions to generate heating, mainly curbing on-site heat generation from primary energy sources (fossil fuels). The firms reduced their natural gas consumption between 21% and 33% and oil consumption between 56% and 48% during Phase II. Again, their findings are statistically significant at 5% level. However, there was no change in their electricity consumption.
[1] Dan Fullerton and Robert Stavins. 1998. “How economists see the environment.” Nature, 395: 433–434. Retrieved from https://www.nature.com/articles/26606 on October 29, 2018
[2] Geoffrey Heal. 2007. “A Celebration of Environmental and Resource Economics”
[3] Ellerman A. D., Marcantonini, C., and Zaklan A. (December 3, 2015) “The European Union Emissions Trading System: Ten Years and Counting”
[4] This information is derived from the class notes
[5] This information is derived from the class notes
[6] This information is derived from the class notes
[7] European Commission Website https://ec.europa.eu/clima/policies/ets_en
[8] Ellerman A. D., Marcantonini, C., and Zaklan A. (December 3, 2015) “The European Union Emissions Trading System: Ten Years and Counting” (p.91)
[9] While ninety -five percent of allowances were allocated freely during Phase I; ninety percent of them were allocated freely during Phase II.
[10] idem
[11] Ellerman A. D., Marcantonini, C., and Zaklan A. (December 3, 2015) “The European Union Emissions Trading System: Ten Years and Counting”
[12] idem p.103
[13] Hintermann, B. (25 October 2009) “Allowance price drivers in the first phase of the EU ETS” in the Journal of Environmental Economics and Management
[14] In more detail, he sets up the model according to a function of temperatures across Europe, reservoir level changes in the Nordic countries, precipitation in non-Nordic countries, fuel prices, the Financial Times Stock Exchange (FTSE) Eurotop 100 (a tradable index representing the 100 most highly capitalized blue-chip companies in Europe).
[15] Hintermann, B. (2009) p.51
[16] Idem p 53-55
[17] Creti,A., Jouvet,P.A., Mignon, V. (29 November 2011) “Carbon price drivers: Phase I versus Phase II equilibrium?”
[18] Creti at all. (2011) p.332
[19] Koch N., Fuss S., Grosjean, G., Edenhofer. O. (11 July 2014) “Causes of the EU ETS price drop: Recession, CDM, renewable policies or a bit of everything? —New evidence”
[20] The two models are similar to each other. While in the first model the switch means fuel the theoretical switching (from coal to gas) price effect, in the second model the price changes of gas and coal are written separately. Economyt refers alternatively to the STOXX EUROPE 600 stock index return or the change of the Economic Sentiment Indicator (to avoid multicollinearity). Wind=Solart and Watert capture the growth of electricity production from wind/solar and water, respectively. Finally, CERt-1 refers to the number of issued CERs (in natural logarithm) lagged by one period, taking into account the fact that newly issued CERs may not be surrendered immediately in the EU ETS.
[21] Koch et all (2014) p.681
[22] Koch et all, (2014) p. 682
[23] Kosh et all. p. 676
[24] Retrieved from https://markets.businessinsider.com/commodities/co2-emissionsrechte
[25] Petrick, S and Wagner, U. (March 28, 2014) “The Impact of Carbon Trading on Industry: Evidence from German Manufacturing Firms”
[26] The reasons for investigating is twofold. Firstly, Germany is the biggest economy and also the largest emitter of EU. Secondly, the researchers benefited from large available data sets in this country.
[27] idem
[28] idem p.18
[29] idem
[30] Idem p.19