The EU Emissions Trading Scheme. Allowance Prices, Trade Flows, Competitiveness Effects

The upcoming European Emissions Trading Scheme (ETS) is one of the more controversial climate policy instruments. Predictions about its likely impact and its performance can at present only be made to a certain degree. As long as the National Allocations Plans are not finally settled the overall supply of allowances is not determined. In this paper we will identify key features and key impacts of the EU ETS by scanning the range of likely allocation plans using the simulation model DART. The analysis of the simulation results highlights a number of interesting details in terms of allowance trade flows between member countries, of allowance prices, and in terms of the role of the accession countries in the ETS.


Introduction
When the European Emissions Trading Scheme (ETS) for CO 2 will start in 2005 it will be known as one of the more controversial climate policy instruments. It is designed to achieve an economically efficient reduction of CO 2 from major energy-intensive installations. Currently it covers around 40 000 installations in the European Union which are responsible for roughly 45 percent of all CO 2 -emmisions in the EU. As of today, there is considerable uncertainty about the impact of this trading scheme when it is in full operation and when the commitments of the member states of the EU to the Kyoto-Protocol will need to be met in 2012. Consequently, speculations sprout about winners and losers among the member states, about costs to different sectors within members states as well as about the question as to which member state will be a net-seller and which one a net-buyer of emission allowances. Also, the range of prices for emission allowances is still wide open. In fact, many statements about the likely outcome of the ETS are more based on the desire to further ones commercial interest than on a balanced analysis of the evidence available so far. Existing quantitative simulation studies (Böhringer 2002, Capros et al. 2000 only analyse preliminary scenarios of the ETS that, for example, do not include the accession countries or account for different likely allocation modes. Predictions about the likely impact and the performance of the European ETS depend on the details of the allocation of emission rights within each member state. As long as the National Allocation Plans of the member states are not finally settled the overall supply of CO 2 -emission allowances is not determined. This obviously influences the price level for allowances. In addition, allocation rules that differ between member states will also influence trade flows. These issues of allocating the caps will be discussed below in greater detail. The second uncertainty concerns the fact that the impact of the EU ETS will exercise its full force in 2012. It is therefore necessary to assess the ETS in the light of the EU economy in the future; to be precise we choose 2012. This will be done with the help of the DART-model (Klepper, Peterson, Springer 2003), a computable general equilibrium model calibrated for the enlarged EU.
Nevertheless, it is possible to identify key features and key impacts of the EU ETS by scanning the range of likely allocation plans and by using a simulation analysis with the DART-model. This approach at the moment ignores some institutional details of the ETS such as the possibility for using the flexible mechanisms set out in the Kyoto-Protocol, i.e. Joint Implementation (JI) or the Clean Development Mechanism (CDM) which can potentially offer further inexpensive abatement options. It also ignores intertemporal issues such as banking and borrowing as well as the links to other national trading schemes in Denmark or the UK or to international trading activities. Despite these omissions the analysis highlights a number of interesting details about the EU ETS in terms of allowance trade flows between member countries, of allowance prices, and in terms of the role of the accession countries in the ETS.
In the following we first summarize the background of the EU ETS and the international climate policy commitments of the EU. We then describe the DART-model and the way in which the ETS is implemented in this simulation model. Finally, we discuss the results of the simulation exercises and draw some conclusions.

Scheme
In the Kyoto Protocol from 1997 the EU agreed to cut down their overall GHG  To reach the European commitments at minimal costs a European Emissions trading scheme (ETS) for CO 2 was designed that is at the heart of this paper. 1 The ETS will start in 2005 and all member states of the European Union will be required to impose binding, absolute caps on CO 2 emissions of facilities in energy activities, the production, and processing of ferrous and non-ferrous metals, the mineral industry and the pulp, paper and board production.  Commission 2003) in which a step by step process to develop a NAP is outlined. The paper suggests that the first step should be to establish the share of the total allowable emissions under the Kyoto Protocol that will be allocated to the installations covered by the trading scheme in a top-down economy-wide analysis. In a next step it is then suggested to collect data from the single installations and companies in a bottom up approach. The allocation of permits to each individual sector is finally determined based on current, historical or average emissions for a certain year. We will talk in more detail about possible allocation modes in section 3.2.
We will ignore in the following the problem of non-CO 2 gases and focus on the question of reductions inside the ETS versus reductions outside and discuss the implications of the different allocation approaches. In addition, we will look at the role of the accession countries. Bulgaria, the Check Rebublic, Hungary, Poland, Slovenia, Slovakia and Rumania , , Malta and Zyprus as well as the three Baltic States will become official members of the EU. 2 As EU members, these countries will also enter the EU ETS. This will influence the costs of achieving the European Kyoto target in two ways. First, Eastern Europe offers abatement opportunities that are much cheaper than in the current EU.
Second, due to the economic recession in the 90ies, Eastern Europe's emissions are now below their Kyoto target. Selling some of their excess emission allowances (called hot-air) in the European ETS will further drive down the permit price and the economic costs in Western Europe.

Simulating the Effects of the EU Emissions Trading Scheme
An assessment of the likely allocation and welfare effects of the ETS requires at least two modelling steps. The first consists of the setting up of an appropriate economic model with which the European economy can be simulated for the time in which the trading scheme will be in full force. The second step involves the design of policy scenarios which are likely to arise between today and the time at which the Kyoto-Commitments are to be met.
As a simulation tool we use the DART-model (Klepper, Peterson, Springer 2003) which will be shortly characterized. We then derive the emission caps for the different member states that need to be met by 2012.
2 Except for the candidate countries Bulgaria and Rumania these countries will join the EU in Mai 2004. For Bulgaraia and Rumania accession to the EU is scheduled for 2007, the beginning of the second trading period of the ETS.

The DART-Model
The DART (Dynamic Applied Regional Trade) Model is a multi-region, multisector recursive dynamic CGE-model of the world economy. For the simulation of the European ETS it is calibrated to an aggregation of 16 regions. Table 1 illustrates the 9 countries or group of countries of the EU including the accession countries of Eastern Europe and the other 7 world regions. In each region or country the economy is disaggregated into 12 sectors. Four of these sectors participate in the ETS. Although there is no perfect match between the installations subject to the ETS and the sectoral structure of DART, we believe it to be sufficiently close. In addition, it also covers about 45 percent of the CO 2 -emissions. The economy in each region is modelled as a competitive economy with flexible prices and market clearing. There exist three types of agents: a representative consumer, a representative producer in each sector, and regional governments. All regions are connected through bilateral trade flows.
The DART-model has a recursive-dynamic structure solving for a sequence of static one-period equilibria. The major exogenous drivers are the rate of productivity growth, the savings rate, the rate of change of the population, and the change in human capital.
The model is calibrated to the GTAP5 data base that represents production and trade data for 1997. The elasticities of substitution for the energy goods coal, gas, and crude oil are calibrated in such a way as to reproduce the mission projections of the EIA (EIA 2002). 3

Integration of Policy Scenarios in DART
The simulation of the ETS requires first a determination of the emission caps for the EU member states. Table 3 shows the Kyoto targets for each region or country based on the EU Burden-Sharing Agreement to the Kyoto-Protocol (also see Figure 1). 4 The cap on country groups is the emission weighted average. The first column represents the percentage reduction required relative to 1990. The second column is derived from the business-as-usual (BAU) run of DART up to 2012 and represents the reduction required in 2012 relative to the BAU in 2012. determine within their National Allocation Plans (NAP) which proportion of the emission reduction is to be supplied by those sectors participating within the ETS, and which proportion is supplied from the rest of the economy. The Commission of the EU suggests three modes of allocating targets in its nonpaper (European Commission 2003): • The "historical emissions approach" (HIS) • The "forecasting approach" (FUT) • The "least cost approach" (LC).
In Finally, the least cost approach tries to take into account the fact that CO 2abatement activities carry substantially different costs in different sectors.
From an efficiency point of view this would not matter if all emission sources were to participate in the trading scheme. But since abatement costs will equalize only within the ETS, there is a danger that the historical and the forecasting approach may lead to strong differences in marginal abatement costs between the sectors within ETS and those outside the ETS. The least cost approach tries to take account of this inefficiency by dividing the cap 4 ACC does not participate in the burden-sharing.
between ETS and Non-ETS sectors in such a way that the different abatement cost levels are recognized. Hence, the least-cost approach allocates relatively few allowances to sectors with low abatement costs.
The   ignore the hot-air in most cases and assume that the emission targets are set at the business-as-usual level. If the hot-air is included we assume that all hot-air is allocated to the trading sectors. We also concentrate in the following on the "least-cost" allocation rule for reasons that will be apparent in the discussion of the results in section 4.1.

Simulation Results
In this section we present the results from simulating the scenarios described in the previous section. We present the results of the DART-model for the year 2012 when the ETS is in full force and the Kyoto-targets under the EU burdensharing agreement need to be met. We first report and discuss the results for the allowance prices. Then we show the trade in allowances across the EU.
Finally, we take a look at the changes in sectoral output and the expected competitiveness and welfare effects.

Allowance Prices
One of the major outcomes of the EU ETS that will determine its allocation effects is a uniform allowance price, i.e. a price on CO 2 , throughout the EU.

Trade in Emission Permits
Abatement costs for CO 2 vary not only within a country but to an even larger degree across countries. As can be seen in Figure 5,   The ETS of the EU turns out to lead to a rather lopsided affair. The accession countries will export allowances to all other member states. Figure 6 illustrates the amount of allowances traded within the ETS. The only exporters are the accession countries (ACC). The overall amount of allowances available in the ETS under the "least-cost-principle" and without the inclusion of hot-air is 1140 Mt CO 2 . Out of this 130 Mt CO 2 will be net exports from ACC, i.e. net trade amounts to roughly 11 percent of emissions in the ETS. The accession countries with allocated allowances amounting to their business-as-usual emissions have available 480 Mt CO 2 of which they export 27 percent. These exports go predominantly to the large countries UK, Germany, and France but also to the Benelux Countries. However, relative to their domestic endowments Austria and Ireland combined will be the largest importers by having 33 percent of their consumed allowances imported. The Benelux Countries come in second with an import quota of 27 percent. In comparison France and Germany import around 6 percent of the emission allowances consumed and the southern regions (SEU) and the Scandinavian member states (SCA) import only around 3.5 percent. Figure 6 also shows that the trade in allowances will be dominated by the electricity sector (EGW). With the exception of France because of its large share of nuclear energy in electricity consumption, more than 75 percent of exports and imports will come from and go to the electricity sector. The rest is dominated by imports from the iron, metal and steel sector (IMS) with pulp and paper products (PPP) and refined oil products (OIL) almost negligible.
There have been speculations about the likely trade structure that would emerge without the accession countries. Figure 7 illustrates this case. Since the low cost option from the accession countries is not available the above mentioned group of countries with the low abatement costs among the EU-15

Competitiveness Effects
A major concern of policy makers and industry is that the ETS will have Finally, it is not true, that the competitiveness of sectors covered by the ETS is affected more than the competitiveness of the sectors outside.
The cross-country flows of CO 2 -allowances from the accession countries towards the West indicate that the ETS will in the first place allow the energyintensive installations within the trading scheme to reduce emissions and consequently production to a lesser degree than under a unilateral climate policy scenario (UNI) in which the EU burden-sharing targets need to be met independently. Simulations with the DART-model show that this is the case but it also carries over to sectors not in the ETS. Figure    Finally, Figure 8 shows that the sectors inside the ETS clearly gain from an emissions trading scheme and are thus affected less from climate policy than the sectors outside the ETS. There are three reasons for this (1) Competitiveness effects depend on the exposure of a sector to the world market. Some of the sectors most exposed to the world market such as the chemical sector are outside the ETS. The detailed sectoral data show indeed that in the unilateral scenario the chemical sector suffers more than the IMS and PPP sector inside the ETS.
(2) The sectors outside the ETS are indirectly affected by the emission restrictions inside the ETS as well. In another paper Peterson (2003) shows that these indirect effects that originate from changes in gross energy prices and demand or from prices of intermediate inputs in some cases even dominate the direct effects from the ETS or the other climate policies.
(3) Reaching the Kyoto targets requires severe reductions outside the ETS as well. As shown already in section 4.1, taxes that are associated with these reductions are much higher than the allowance price. As a result, the sectors outside the ETS are affected more strongly than the sectors inside.
Finally it should be noted, that the strength of the effects differs between individual countries. Some of the main factors that influence this strength are discussed in the next section. In addition, the differences in the energy intensity as e.g. described for the chemical sector in the Benelux countries do play a role. For more details see also Peterson (2003).

The Welfare Costs of Different EU Climate Strategies
The main goal of the EU emissions trading scheme is to reduce the welfare costs of meeting the European emission targets. Figure 9 shows the aggregated EU welfare changes relative to the BAU scenario in the different trading scenarios compared to a scenario where the individual commitments are reached unilaterally by a uniform, country specific CO 2 tax (UNI) .  However, this gain is due to the fact that the overall amount of emissions is substantially higher than in the trading scheme without hot-air.
Turning to the economic costs for the individual EU member countries, these can differ considerably. Figure 10 shows the welfare changes across countries for the UNI and the LC scenario. Again, the welfare cost under LC is the light grey bar and under UNI it is the sum of the light and dark grey bars. trading. The degree to which they gain depends on two factors: • The strictness of the national Kyoto-target relative to the business-asusual emissions, and • the differences in abatement costs across different member states.
Both of these factors are illustrated by the implicit CO 2 -taxes necessary to achieve the Kyoto-targets unilaterally (see Figure 5 in section 4.2). The largest gains from ETS accrue to the Benelux countries (BEN) and Austria and Ireland (REU) because they experience the largest difference between allowance price and unilateral tax rate.
France experiences no gains from trading in the ETS although it has the same implicit unilateral CO 2 -tax than Germany which can lower its welfare costs from 1.2 percent to 1.0 percent. This is due to the fact that France is not trading many allowances because of its low emission intensity in the electricity sector.
Hence, it can not reap large gains from trading as much of the emission reduction will need to take place outside the ETS. to the rest of the EU because their production cost including the emission constraint rise more than those in ACC. The introduction of the trading scheme raises allowance prices in the ACC such that the comparative advantage of the energy-intensive sectors is reduced. This trade effect can not be compensated by the income received from exporting allowances as the volume is too small. However, if the ACC would use some of their hot-air, they could reduce allowance prices thus reducing the burden to their ETS sectors and at the same time increase the income from exporting allowances. In fact, with a supply of 25 percent of the hot-air available the welfare loss could be reduced to zero. Larger amounts of hot-air would even lead to a welfare gain for ACC of 0.8 percent relative to BAU:

Summary and Conclusions
The