Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser
Impact assessment of Proposal for a Directive on the limitation of emissions from medium combustion plants – National impact assessment compared to the European impact estimate ⁎
Ondřej Vojáčeka, , Ladislav Sobotkab, Jan Macháčc, Miroslav Žilkaa a b c
Czech Technical University, Czech Republic University of Economics, Prague, Czech Republic J. E. Purkyne University in Usti nad Labem, Czech Republic
A R T I C L E I N F O
A BS T RAC T
Keywords: Medium Combustion Plants Directive Impact assessment Emission reduction Emission of SOx NOx PM EU environmental policy
The December 2013 Medium Combustion Plants Directive (MCP Directive) proposal was evaluated by the national governments. In the Czech Republic, there are 6710 plants affected by this Directive, which is about 4.6% of the total of 143,000 relevant European plants. The paper introduces our approach of policy impact assessment called SimTool. The costs estimated for the European Commission in the background study (AMEC, 2014) are assumed to reach EUR 5.9 million for the Czech Republic for the preferred scenario by the European Commission. Further presented national impact assessment estimates the induced annual costs of the proposal at EUR 61 million, which is about 10 times greater than the European impact assessment estimate. As part of the national impact assessment, the different fuel categories had to be analyzed separately due to their specific features and different options for achieving the emission limit target values. During the impact assessment, a survey was made in order to determine the source operators’ preferences and responses to the potential adoption of the MCP Directive. Based on the analysis of data from the operators and consultations with experts about the different technologies, technical options for achievement of the proposed emission limits, including an estimate of the operating and investment costs, were proposed. The paper concentrates on the Czech impact assessment approach and discusses the reasons of the discrepancy between the European impact assessment and the Czech version. We argue that the inaccuracies of the European impact assessments are given by usage of the general abatement cost curves in the models which do not reflect the reality sufficiently. This paper states an argument for the necessity to carry out analysis at the local level.
1. Introduction Recently, there has been an increasing demand for clean air within the EU area as well as on the global scale. The European Union is striving for a unification of the European legislation for different polluters in an effort to meet the commitments to protect global climate and achieve a cleaner air quality. In the first phase, mostly large combustion sources with a capacity over 50 MW were concerned. Now, the EU's attention has shifted to medium and small combustion sources. The regulatory attention on emission sources smaller than
large combustion plants is actually global, as these efforts prove to be effective also outside the EU [1]. The proposal to introduce a single “Directive on the limitation of emissions of certain pollutants into the air from medium combustion plants with an installed capacity between 1 and 50 MW” (hereinafter, the MCP Directive [2],) is part of the package of measures to clean up Europe's air. As part of the negotiation process, the Directive proposal has undergone an impact assessment at the national level and the level of regulated substances. The starting point for the assessment was the AMEC (2014) [3] study, which assessed several different scenarios. The
Abbreviations: capex, capital expenditure; CZK, Czech crowns; EL, emission levels; EU, European Union; EUR, Euros; GAMS, General Algebraic Modeling System; GJ, gigajoule; IED, Industrial Emissions Directive; MCP, medium combustion plants; MS, member states; MW, megawatt; MWt, megawatt (thermal); NOx, nitrogen oxides; opex, operational costs; PM, particulate matter; PRIMES, Price-Induced Market Equilibrium System; RAINS, Regional Air Pollution Information and Simulation; RIA, regulatory impact assessment; SOx, sulfur oxides; SO2, sulfur dioxide ⁎ Corresponding author. E-mail addresses:
[email protected] (O. Vojáček),
[email protected] (L. Sobotka),
[email protected] (J. Macháč),
[email protected] (M. Žilka). http://dx.doi.org/10.1016/j.rser.2017.06.119 Received 24 December 2016; Received in revised form 25 June 2017; Accepted 26 June 2017 1364-0321/ © 2017 Elsevier Ltd. All rights reserved.
Please cite this article as: Vojacek, O., Renewable and Sustainable Energy Reviews (2017), http://dx.doi.org/10.1016/j.rser.2017.06.119
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
Table 1 Target emission levels (mg/Nm3) and costs of achieving proposed emission limits in the Czech Republic under different options (million EUR). Source: Own adaptation from [13] (pp. 165–175)
SO2 NOX PM Total a
Business as usual
IED 50–100 MW
Emission levels
Emission levels
2559–3206 978–1226 376 –
400 300 30 –
Total costs 4.2 16.5 2.7 23.3
Most Stringent MS
Gothenburg
Emission levels
Emission levels
Total costs
200 100 5 –
5.2 25.4 3.5 34
Holds only for new plants, existing plants follow Gothenburg.
Diagram 1. Illustration of impact assessment of individual plants. Source: Own approach
2
400 680–800 30 –
SULES Total costs 3.4 0.3 2.3 5.9
Emission levels a
200 100a 5a –
Total costs 3.6 3.7 2.3 9.6
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
Diagram 2. Illustration of Impact assessment at the level of individual plants. Source: Own proposal
used continuously in the Czech Republic for almost 8 years for different impact analyses [5,8,9], and is known as the SimTool approach and the SimTool model. The paper demonstrates the impact of the MCP Combustion Plants Directive proposal on the Czech Republic, containing an estimate of costs and other impacts of the Directive on the source base of the Czech Republic. An introduction and description of the impact assessment methodology called SimTool is made in the first chapter; in the second chapter the microeconomic model SimTool is briefly described. The results of the impact assessment of the MPC Directive application in the Czech Republic are introduced and discussed within the context of the official European impact assessment.
scenarios differ in terms of the stringency of the emission levels for particular pollutants and the corresponding cost. In the “IED 50– 100 MW” scenario the emission levels are set to be equal to the ones required by IED. The “Most Stringent MS” scenario sets the emission levels to match the most stringent emission levels in member states’ national legislation. The “Gothenburg” scenario uses NOX emission levels set out in the amended Gothenburg Protocol and follows “IED 50–100 MW” in the case of SO2 and PM. The “SULES” scenario copies Gothenburg for existing plants, but sets the emission levels to equal the “Most Stringent MS” scenario for all new plants. See Table 1 for the emission levels and the cost estimates for the particular scenarios. It is evident from the table that the assumed annual costs of the Directive are between 5.9 and 23.33 million Czech Crowns (CZK) depending on the scenario, the average annual costs being CZK 18.2 million, whereas the scenario preferred by the European Commission is shown in the column Gothenburg, with annual costs totaling CZK 5.9 million [4]. This amount appears to be greatly undervalued with respect to the stringency of some of the newly proposed emission limits and experience with the preceding regulation for sources with a capacity over 50 MW of thermal input. Clearly, the results of the impact assessment vary greatly depending on the valuation tool chosen, and there are many tools developed for environmental policy impact assessment. A large share of which is developed for a specific policy analysis only; e.g., about 40% of the designed project tools within the European Framework Programs did not fit into any of the categories [5]. Secondly, within a broad scale of options among approaches toward impact assessment, there is a call for the use of more evidence-based tools to enhance the quality of the impact assessment work [6], as the general impact assessments computed via macro models are based on strong underlying assumptions and they are usually not very transparent (as e.g. PRIMES model, which is privately owned and its results can not be replicated). To the authors’ awareness, no standardized microeconomic analysis has been formulated in previous research. This corresponds to the review of [7] the small and medium enterprises regulation case studies, which identically showed a lack of standardized means to evaluate industrial-energy end-use policies. This paper strives to introduce such methodology, which has been
2. The methodology (SimTool approach) This chapter describes the application of the SimTool approach to the impacts of the MPC Directive. The approach is general and it is described in relation to the above mentioned Directive (see Diagram 1 below). The first step of the impact assessment is to identify the potentially endangered companies. The split into groups according to intensity of impact can be made under various conditions. One of the opportunities for the combustion plant sector is to carry out an analysis of the deviation between the current value of the regulated parameters and the target values. Based on such an analysis, all plants can be divided into two groups: plants complying with the target values and those not complying with them. Further classification is made according to the pollutant type: this case is depicted in Diagram 1, where “Group IV” stands for plants that comply with the requirements set by the MCP Directive and the other groups represent plants that do not comply. The diagram also depicts a way of classification of the intensity of the impact, where “Group I” represents sources not complying for all 3 here considered pollutants, “Group II” does not comply with limits for 2 pollutants and “Group III” for only 1 pollutant. After categorization of the plants, we need to select representative plants for each group and get data needed for the impact assessment simulation on these representative plants (See the following chapter “SimTool microeconomic model” for details). The modeling of the economic impact at the level of an individual company is depicted in Diagram 2. Initially, a base scenario is defined, where the economy of the plant 3
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
Diagram 3. Schematic illustration of impact assessment in cases where regulation leads to substantial revision of business conditions. Source: Own proposal
is simulated without the regulation considered (see “Basic scenario without evaluating regulation” depicted in Diagram 2). During this phase, we have to consider mainly (i) the external environment of the plant (price development and other inputs, price development of the outputs and other parameters related to outputs – including waste and emissions, price development and the amount of substitutes and complements), and (ii) data about economy of the plant, balance sheets, profit and loss statements, company cash flow, including information about factors that are important from the point of view of the current state and the future outlook of the business activities. In some cases, a simplified approach proved to be sufficient, when only the balance sheets, profit and loss statements for the recent years (3–5 years) plus investment plans and comments from the management are used. This approach is valid if a company profitability margin can be used as a sufficient criterion of the company's success and subsequently change of the outputs can be modeled in order to achieve the externally set profit margin. Within the heat generation sector, it was possible to apply such a simplified approach and it was not always necessary to also project assets, which subsequently allow calculation of additional performance indicators (such as Return on Equity, Return on Assets). Among external conditions, there are also all currently valid regulations that have an impact on the plant (either directly or through the impact on the price or other parameters of the inputs or outputs). Furthermore, we need to postulate certain assumptions, mainly set the fix and the variable part of the individual items within the accounting data and their expected change in time. Based on the information given, it is possible to carry out a projection of the economy of the plant in the base/reference scenario: a scenario without the regulation considered. Defining the scenario with the regulation involves several steps (see the “Technological change not necessary” and Technological change is necessary” parts of Diagram 2). Firstly, we need to find out whether it is possible to meet the requirements without changing the technical parameters of the plant. This is, for example, the case of a change of the fees for emitted pollutants or a change of the free allocation of emission allowances. Such regulatory changes can indeed trigger some changes in the decision-making of the plants, as the plant operator will strive to minimize the economic impact (e.g., by increasing the share of biomass
in the fuel mix); however, these changes are not necessary to meet the requirements. The operator can always choose just to pay the higher costs of emissions or allowances. The second option (as in Diagram 2 “external conditions + shock with no technological change”) is a situation where regulation changes the requirements on plant operation and implies some technological changes, e.g., strengthening of emission limit, which implies either a shift toward fluid boiler technology or at least end-of-the-pipe technology (e.g., implementing desulfurization). If the regulation induces such requirements, managements of the selected plants should identify potential reactions to the new situation mainly from the point of view of technology. Should more realistic options exist to meet the regulatory criteria, it is necessary to describe these and economically assess them (as in Diagram 2 “Economic parameters of the technological change 1 – 3”) and subsequently implement them in the scenario with the regulation. Results of the scenarios are also described by a set of indicators, compared to the reference one and interpreted in the context of the external environment of the plant (see Diagram 2 “Reaction on the external conditions”). This part is crucial, as the plant does not exist in isolation but its performance depends on market interactions. Thus, we evaluate whether the changes in the company performance are in line with surviving on the market where the company operates. In other words, if the necessary costly reconstruction makes the plant noncompetitive, this scenario has to be excluded from the technological options of the reaction to the regulation. If all options are excluded one by one, termination of operation is the last one available (or transition to another product) (as in Diagram 2 “Serious impact, possible production change, possible production closure”). On the other hand, if one or several assessed reactions to the change in the regulatory conditions implies such changes in economic parameters that allow company competitiveness we can consider these reactions to be the likely costs of the regulation at the level of the individual plant (See “Optimal scenario of company reaction” in Diagram 2) and carry out an aggregation/extrapolation of the costs over the whole group of plants in the given emission class (see Diagram 1 “Total regulatory costs for group I to III”). A sum of the costs in the individual groups returns the total costs of the assessed regulation (see Diagram 1 for “Total costs of the regulation”). 4
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
corporate financial analysis. In the modeling phase, the attention focuses on the simulation of accounting reports. The simulations are designed in a way to reflect the future most accurately. Thus, they are based on real accounting data and plant-specific information to minimize the externally set assumptions. The model is set, based on the computational solvers for models or general equilibria. The process of the simulation is such that initial parameters of the simulation are computed based on the internal relations and set variables. Subsequently, the external shock is defined and also the variables are set, so that the final values lead to the final equilibrium. One of the options usually considered is where the price of the outputs is set in a way that the gross and the profits are balanced for each scenario, which enables achievement of the required profitability. The model is based on an input-output methodology, where each unit of production requires a proportionally equal amount of inputs. Technological change (e.g., reconstruction) then changes the proportional amount of inputs needed per unit of production. The advantage of such an approach is a significant decrease in simulation parameters; however, there is a disadvantage to the combined generation in the form of assigning the costs to multiple outputs.
If no reaction to the change in the regulatory conditions is economically viable and regulation thus leads to a substantial revision of the conditions of the business, then we need to analyze the cases more deeply (see Diagram 3). The reason is that we cannot consider the costs of any of the assessed options of the reaction to the new regulatory conditions as the costs of the regulation at the level of the individual plants (so-called compliance costs), are crucial cost-wise. Thus, the analysis expands and examination of the group of plants (according to the classification of the intensity of the impact defined within the Diagram 1) must be carried out (see Diagram 3 for “Group I”). This also happened with the MCP Directive impact assessment and several plants were thus randomly selected for each fuel group and their reactions on the new situation are observed (see Diagram 3 “Reaction A” to “Reaction X”). The answers are to be analyzed and maybe some behavioral patterns or systematic reactions can be identified that would be added to the extrapolation of the plant reactions and that would further precise such extrapolation. This would be, in other words, modeling reaction of the plants as a dependent variable of the observed parameters. After classification of the expected reaction of the plants, costs of individual reactions are calculated. Here, it is necessary to calculate costs of the transition to the fluid technology of combustion, transition to biomass boilers, gas boilers and central heating system decentralization. Based on these costs and the identified expected reactions of the plant operators to the evaluated regulation, the cost of the reaction is extrapolated to the whole set of plants in the same category according to the intensity of the impact. Subsequently, the total costs of the regulation are derived. This procedure is repeated for all the groups of plants with significant impacts.
3.1. Medium combustion sources and current state of emission production The results of the methodology described above applies to the Medium Combustion Plants Directive in the Czech Republic are shown in the following chapter. There are 6710 medium combustion sources in the Czech Republic, which is about 4.6% of the total of 143,000 medium combustion sources in the EU [10,11]. Sources combusting natural gas comprises the largest portion (79%). Besides this dominant category, a great number of coal (5%) and liquid fuel sources (8%) are in operation. Most of the medium combustion sources in the capacity range of 1– 50 MW are utilized for the heating of households, the business sector and public institutions: just fewer than 4500 sources, making up 66.6% of the total. Said sources produce approximately 66,925,000 GJ of heat a year; 74% of the heat is made from natural gas. Coal is the second most common fuel with a production of 9.8% of the heat. The table below shows the quantities of heat produced by the different fuel categories. The share of the heat production is about 60% (Table 2). Heat production is associated with emission production. The proposal for the MCP Directive focuses chiefly on PM, SO2 and NOX emissions and their limits. These limits will be reduced in 2018 due to national legislation; the EU legislation expects introduction of limits starting in 2025. At present, the Czech Republic produces 702 t of PM, 6380 t of sulfur and 6650 t of NOX annually from the basic source categories by fuel type [9]. The emission limits are derived from the permissible emission concentrations. Greening efforts lead to a constant pressure to reduce
3. SimTool microeconomic model (general description) The SimTool model is built in the environment of the GAMS modeling language (General Algebraic Modeling System) and it joins vast opportunities of its language and microeconomic theory of the firm based on accounting principles [9]. The model provides an adjustable environment for creation of projections and of scenarios of future economic development, assets and cash flow at the level of individual companies. Generally, it entails that individual scenarios which can be analyzed by SimTool differ mainly in the different investment plans, depreciation, subsidies and borrowings, changes in the fuel mix and consumption of fuels, efficiency of boilers and turbines, changes in emission fees, emission allowance costs, costs of materials for desulfurization and denitrification). Output values of the default scenario and scenarios with regulation are then compared. Thus, this approach is based on the financial analysis of companies, where only the marginal and shadow values in the selected company parameters in the given time horizon are focused on. Thus, it is an approach that is similar, up to a certain point, to analysis and assessment of investment projects in Table 2 Heat production by fuel type. Source: Own adaptation from REZZO database [11] (2012 data) Category
Total annual heat production in GJ for the category 1–50 MW
Share in total heat production
Within that, produced in the sector “combustion plants for electricity and heat production for households, businesses and public institutions”
Share in production of the sector “combustion plants for electricity and heat production for households, businesses and public institutions”
Coal Biomass Liquid fuels Natural gas Biogas Total
6,583,058 5,274,396 2,300,800 49,696,455 3,070,786 66,925,495
9.80% 7.90% 3.40% 74.30% 4.60% 100%
5,387,994 4,178,992 1,001,035 27,734,292 613,415 38,915,728
13.90% 10.70% 2.60% 71.30% 1.60% 100%
5
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
shown in Table 4. These 5 micromodels represent the basic model situations in the sector of screened coal sources. They indicate that the average increase to the heat price for this category is EUR 19.3 per GJ. It turned out during the source analysis that many of the sources are very old, with innovation being very costly. Source replacement would be the only economically viable option. Referring to the described modeling approach, a survey was made in order to determine the source operators’ preferences and responses to the potential adoption of the MCP Directive. Around 210 sources non-compliant with the proposed emission limits were approached. Based on the analysis of data from the operators and consultations with experts on the different technologies, technical options for achievement of the proposed emission limits, including an estimate of the operating and investment costs, were proposed. The cost estimates were further refined based on data acquired from the empirical survey and then used to model the costs of achieving the emission limits for the entire source base. Table 5 shows the costs of proposal compliance of the screened coal category. The columns are based on the empirical survey results and represent options for operators’ behavior. The majority of the operators will deal with the emission limits by shifting to gas. The introduction of the new emission limits in this category will require capital investment costs amounting to EUR 100 million and annual operating costs of EUR 8.0 million. It turned out when the source operators were approached that liquid fuel sources are already in a decline at present, with operators gradually replacing them with other fuel sources. Once effective, Decree no. 415/2012 Coll. will lead to a reduction of limits, resulting in the shutdown of almost all the liquid fuel sources. A further reduction to the emission limits based on the MCP Directive will therefore not have a major impact on this category. Thus, the adoption of the Directive will have the biggest impact on the coal, biomass and biogas source categories. The impacts on other sources were calculated analogously to the screened coal sources. The total investment costs of achieving the emission limits as per the MCP Directive proposal exceed EUR 407 million. The structure of the investment costs by the fuel type is shown in Table 6. The investment costs do not include the costs of sources combusting liquid fuels, as the operation of these sources will be ended as a consequence of national legislation by 2018. Thus, the MCP Directive proposal has no effect on this category of sources. A similar conclusion can be made for sources combusting gaseous fuels, as these sources will have been regulated to the level required by the MCP Directive proposal already based on national legislation. The operating costs of achieving the emission limits as per the MCP Directive proposal are EUR 21.2 million. They comprise of the increase in the operating costs due to the interception of PM and SOx and the costs increasing the heat price in the case of sources shifting from coal to natural gas. Table 7 shows the structure of the operating costs by the fuel type.
Table 3 Emission limit values (mg/Nm³) for medium combustion plants as per the MCP Directive proposal. Source: Own adaptation from [European Commission, 2013) [3]] Pollutant
Solid biomass
Other solid fuels
Liquid fuels other than heavy fuel oil
Heavy fuel oil
Natural gas
Gaseous fuels other than natural gas
SO2 NOX Particulate matter
200 650 30
400 650 30
170 200 30
350 650 30
– 200 –
35 250 –
the limits and operate combustion sources in more environmentally friendly ways. Requirements as per the current EU proposal are shown in Table 3. Comparing the above values with the limits currently in force, emission limits are not set for some of the fuel and emission types in the category of the medium combustion plants, or they are several times higher (by up to an order of magnitude in some cases); a similar situation exists comparing the proposed limits with limits in force as of 1 January 2018 pursuant to Decree no. 415/2012 Coll. The adoption of the MCP Directive therefore requires a perceptible decrease in emission concentrations. The regulatory approach via introduction or strengthening emission limits proves to be inexpensive in terms of regulatory costs, transparent and, importantly, also technology-neutral. In comparison with the technology-specific approach, this approach does not lock in particular selected and incentivized technologies. This would in the long term lock out emerging energy technologies and lead to higher total generation costs in the future than if a technology-specific approach was used in the short term [12]. 4. Impacts of MCP Directive on the fuel categories As part of the impact assessment, the different fuel categories had to be analyzed separately due to their specific features and different options for achieving the emission limit target values. This paper presents the entire procedure for the category of sources combusting screened coal. The number of sources in each category that would not comply with the emission limits in question was determined from the REZZO database [11]. In the category of screened coal sources, 370 out of the 385 sources would not comply with the emission limits; the majority of the sources currently non-compliant with the proposed limits have a capacity of 1–5 MW. Afterwards, possible solutions were proposed with respect to technical availability. The SimTool model described above was used for modeling the impact of the MCP Directive requirements on selected non-compliant sources, informing about the costs of achieving the emission limits and their projection in the price of heat. The results of 5 micromodels of screened coal sources are Table 4 Price increase due to compliance of screened coal sources with emission limits. Source: Own analysis
Total heat capacity (MW) Dedusting investment (million EUR) Desulfurization investment (million EUR) Total investment to achieve EL (million EUR) Price increase due to dedusting, incl. variable costs (EUR/GJ) Price increase due to desulfurization, incl. variable costs (EUR/GJ) Total heat price increase (EUR/GJ)
Plant 1
Plant 2
Plant 3
Plant 4
Plant 5
Average
5 0.4 0.9 1.4 4.6 9 13.6
1,55 0.1 0.4 0.5 3.9 5.8 9.7
2,8 0.3 0.4 0.7 2.3 11.3 15.5
3,3 0.3 0.6 0.9 19.9 25.4 45.4
1,5 0.1 0.4 0.5 4.7 7.7 12.4
2,83 0.3 0.5 0.8 7.1 11.9 19.3
6
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
2,909,910 8037 101.6 136,402 22 6.5
363,739 1493 17
727,478 0 –
181,869 – 1.9
The above mentioned yearly costs for the Czech Republic (EUR 62.85 million) are more than ten times higher than the ones estimated by [3] in the scenario preferred by EU (EUR 5.93 million), after annualisation it becomes EUR 58.63 million. The main reasons for the cost differences are as follows. Firstly, there is a different number of plants evaluated in each study. While [3] claim that detailed dataset was provided by the Czech Republic, the total amount of plants used in the study (4991) falls short of the number used in this paper (6710) according to [11]. Using a full sample, costs would increase by 33% considering the plants left out by AMEC are distributed identically in the sample. More importantly, both approaches differ in assigning measures to specific plants. The study [3] assigns the measures based on abatement matrices, which select measures for each plant that ensure satisfying the respective emission limit. This is to highly extend a general approach not reflecting the market environment of the companies and the need to stay competitive after the automatically applied measure is paid. However, this is a very common feature of the impact assessment approaches applied within the EU. In the SimTool approach the influence of the application of the respective measures to fulfill the regulation of the company is examined in the context of the respective market, possible substitutes for the final users and their price. When there is a serious impact to the companies’ competitiveness and business rent ability it is needed to carry out the research of the reactions on the representative sample of individual plants to find out their behavior if the regulation enters into force. Using this approach may increase the total compliance costs, because the plants do not always pick the cheapest solution (see the process of reaction selection depicted in Diagrams 2 and 3), instead, they pick a long-term solution, which may be different from the one used by [3]. Our approach reflects a real decision making process of the companies on the market. This approach has a possible flaw as the plants that could not decide on a strategy were assigned an average cost of the regulation compliance of the decided operators. Also our analysis allows the plants to shut down permanently and the costs are in such cases calculated as a replacement of the installed capacity by the decentralized heating systems with their market cost of installations and operation. The decentralization possibility with its cost is omitted in [3]. Another difference can be found in sets of measures used in the studies as they are not identical. Some abatement measures were not included in [3]. Crucially, implementation of pulverized coal-fired boiler was left out, which was however in the Czech analysis one of the most appealing options based on the answers received from the operators. The abatement measures included in both analyses also differ in costs as [3] relies on information from literature, while cost of the abatement measures used in the SimTool approach are based on several data sources: companies developing and installing the technologies, technical consultations with university experts and, where possible, from previous installations. Real offers proposed to the individual operators and data from the companies developing and installing the emission reduction measures. For example, annualized costs of the implementation of a fabric filter is valued at EUR 2768 – 6617 in [3] for plants with capacity between 1– 5 MW. Recalculating the upper bound estimate to the net present value of the investment gives EUR 73,570 which is approximately 3.5 lower than the offer made in a case study to a Czech 1.55 MW plant. This comparison was replicated on various abatement measures and the results show that the real investment is 2–20 times more expensive than the suggested values expected by AMEC study and subsequently by the EU impact assessment.
363,739 167 16 272,804 130 12
863,880 6400 48.3
266 17 67 33 12 33 25
79
5. Differences in cost estimates between AMEC and SimTool
Number of sources non-compliant with emission limits Amount of heat produced (GJ) Annual operating costs (thousand EUR) Investment costs (million CZK)
Effort to retain source by adding equipment Shift to fluidized bed boiler
Table 5 Costs of complying with proposed EL, screened coal category. Source: Own analysis.
Shift to gas after 2018
Shift to biomass
Unable to estimate next steps at the moment (do not know)
End of operation
Replace with a source under 0.5 MW
Total
O. Vojáček et al.
7
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
Table 6 Total investment costs of achieving proposed emission limits as per Directive proposal (million EUR). Source: Own analysis. Fuel type
Screened coal
Pulverized coal
Biomass
Liquid fuels
Biogas
Total
Total investment costs in category (million EUR)
102
225
54
–
36
417
along with Estonia, Slovakia, Malta and the Netherlands, provided the largest amount of data required for the development of the impact study [13]. Similar results derive from estimates for Finland [14], where the European impact assessment study presented PM compliance costs of EUR 1.3 million annually per ton, while a national impact assessment arrived at the result of EUR 14 million annually per ton. The comparison can be thus summarized as following (Table 9). These findings are in line with other authors who also favour more evidence-based tools to enhance the quality of the impact assessment [6], especially within the very non-unified environment [5]. The national impact assessment using the SimTool approach and model is detailed and does not consider any automatic modeling procedures as does the AMEC study. The SimTool approach is an intuitive, logical bottom-up approach recognized both by the industry and the regulators. The required data at the level of companies was obtained from regulated companies’ management; the abatement costs were obtained from companies developing and installing the technologies, technical consultations with university experts and, where possible, from previous installations. The AMEC study is based on a database of medium combustion plants in the Czech Republic [15] and on projections using PRIMES 2012 data on fuel consumption per GAINS sector for 2010, 2025 and 2030. The current combustion plants were categorized based on installed capacity and the type of fuel. According to the AMEC study, the Czech Republic provided all information about numbers of combustion plants, their fuel consumption capacities, emissions of key pollutants, existing coverage and directly associated activities and legislative requirements. Only the data about technology types used was not complete. Installation plant categories were set according to the data on combustion. An abatement matrix was developed as part of the study which contains details of applicable abatement measures for each plant category (size and type) and for reduction of each pollutant (SO2, NOx and PM). The costs were calculated for each abatement measure in one of the installed capacity intervals. These costs included both capex and opex. The data about medium combustion plants was merged with abatement matrices and thus compliance costs were estimated. The information on the most feasible measures and costs is based on Vito [16], AEA [17] and AMEC [13]. A matrix of possible measures was generated based on a literary survey. A wide range of costs of the measures was defined based on the survey of references. The most suitable measures, which achieve the required reduction, were selected automatically from the matrix of measures. If the difference between the existing and required emission levels was smaller than 10%, no costs were taken into account; in this case, the AMEC study expects that achievement of the limits is feasible by only a small adjustment to the technology. If the reduction requirement is greater than 10%, data on costs of the change were used as the range minimum. The costs of the change were expressed in the form of annualized costs, including both investment and operating
6. Emission savings The total amount of emissions from 1 to 50 MWt sources in the Czech Republic is 6379 t of SO2/year, 6651 t of NOx/year, and 702 t of PM/year, representing shares in the total emissions in the Czech Republic of 4.13%, 3.15% and 1.17%, respectively. The figures indicate that even a total elimination of emissions from 1 to 50 MWt sources would probably only have marginal impacts on the total pollution load in the Czech Republic. The amount of emission savings can be specified with respect to the changes associated with compliance with the proposed emission limits for the different fuel categories. No decrease would occur in the natural gas and liquid fuel categories as a result of the adoption of the proposed Directive: the sources would have been replaced due to the more stringent national legislation that will become effective in 2018. The decrease in the amounts of emissions produced by the other sources non-compliant with the proposed emission limits averages around 65%. The greatest emission decrease compared to the current situation would be for PM from screened coal: the models indicate a decrease in PM of up to 89%. If measures to comply with the emission limits as per the MCP Directive proposal are implemented, the greatest reductions will concern PM and SOx. The total sum of all the 4 source categories would be a decrease of 584 t of PM/year, 4795 t of SOx/year, and 519 t of NOx/year. Table 8 shows a detailed breakdown by categories.
7. Discussion and conclusions The national impact assessment calculated the total capital investment costs of the proposal for the MCP Directive for sources in the Czech Republic; they were quantified at EUR 409 million, with annual operating costs of EUR 20.8 million. The total annual costs would thus be EUR 61 million. The quantified costs are based on conservative estimates as per the current conditions and the REZZO database [11]; the actual costs would probably be higher. The costs estimated for the European Commission in the background study [3] for different scenarios range between EUR 5.9 and 23.3 million, whereas the scenario preferred by the European Commission assumes annual costs to the Czech Republic of EUR 5.9 million. It follows from an extensive empirical survey carried out as part of the present study among operators of both small and medium combustion sources that the expected costs of the MCP Directive would probably be approximately 10 times higher than those assumed by the European Commission's impact analysis. The capital investment costs correspond to 140,000 EUR/MWt for sources combusting screened coal, and 770,000 EUR/ MWt for sources combusting pulverized coal. The costs are 92,000 EUR/MWt for sources combusting biomass. The cost estimate in the detailed impact study presented in this paper is thus almost ten times higher than the estimate of the official European Commission impact assessment. This conclusion is very serious, particularly with respect to the fact that the Czech Republic,
Table 7 Total annual operating costs of achieving proposed emission limits as per Directive proposal (million EUR). Source: Own analysis Fuel type
Screened coal
Pulverized coal
Biomass
Liquid fuels
Biogas
Total
Total annual operating costs in category (million EUR)
8
6
7.1
–
1.9
21.1
8
Renewable and Sustainable Energy Reviews xxx (xxxx) xxx–xxx
O. Vojáček et al.
Table 8 Summary data on emission saving from sources non-compliant with proposed EL. Source: Own analysis.
Pulverized coal Screened coal Biomass Biogas Total
Current PM emissions, t/ year
New PM emissions, t/year resulting from MCP
PM emission saving, t/ year
Current SOx emissions, t/ year
New SOx emissions, t/year resulting from MCP
SOx emission saving, t/ year
Current NOx emissions, t/ year
New NOx emissions, t/year resulting from MCP
NOx emission saving, t/year
79.5 210.1 484 – 773.6
27.7 23.9 138.1 – 189.7
51.8 186.3 345.8 – 583.9
2643.5 3301.1 250.7 307.5 6502.7
671.7 860.7 117.3 58.3 1708
1971.9 2440.3 133.4 249.2 4794.7
– – 145.6 1013.6 1159.2
– – 94.8 545.7 640.5
– – 50.9 467.8 518.7
Acknowledgements
Table 9 Comparison of cost estimates (mil. EUR). Source: Own analysis.
European impact assessment National assessment Ratio
Czech Republic
Finland
5.9 61 10.3
1.3 14 10.8
This research has been supported by the Ministry of Industry and Trade of the Czech Republic, we are grateful for providing the data from REZZO database. References [1] Zheng S, Masayo W, Yi H. The impacts of provincial energy and environmental policies on air pollution control in China. Renew Sustain Energy Rev 2015;49:386–94. [2] European Commission. Proposal for a Directive of the European Parliament and of the Council on the limitation of emissions of certain pollutants into the air from medium combustion plants; 2013. [3] AMEC. Analysis of the impacts of various options to control emissions from the combustion of fuels in installations with a total rated thermal input below 50 MW; 2014. [4] European Commission. Impact assessment; Proposal for a Directive of the European Parliament and of the Council on the limitation of emissions of certain pollutants into the air from medium combustion plants; 2013. [5] IREAS, e-Academia. Studie ekonomických dopadů splnění emisních limitů podle směrnice 2010/75/EU na menší výrobce tepelné energie, Teplárenské sdružení ČR. Czech; 2011. [6] Backlund AK. Impact assessment in the European Commission – a system with multiple objects. Environ Sci Policy 2009;12:1077–87. [7] Thollander P, Masayo W, Kimura O, Rohdin P. A review of industrial energy and climate policies in Japan and Sweden with emphasis towards SMEs. Renew Sustain Energy Rev 2015;50:504–12. [8] Pur L, Vojáček O, Pícha K. Carbon price and biomass co-burning as a determinant for decision making in green investment. In: Šauer P, Šauerová J, editors. Environmental economics and management. Praha: Nakladatelství a vydavatelství litomyšlského semináře; 2010. [9] Vojáček O, Pur L. Impact of the EU emission trading system on microeconomic level. In: Haas R, Jilkova J, editors. Energy for sustainable development II. Alfa publishing; 2010. p. 91–118. [10] Presentation by Ing. Petr Julínek, TENZA, a.s. Opatření ke splnění požadavků směrnice 2010/75/EC, zákon 201/2012 důsledky pro provozovatele zdrojů; opatření ke splnění požadavků směrnice pro ssz (mcp), příprava legislativy, důsledky pro provozovatele zdrojů, Teplárenské sdružení. Czech. [dataset]; 2014. [11] Data from the REZZO database (CHMI) for small and medium combustion sources. [12] Santana P. Cost-effectiveness as energy policy mechanisms: the paradox of technologyneutral and technology-specific policies in the short and long term. Renew Sustain Energy Rev 2016;58:1216–22. [13] AMEC. Multi Pollutant Measures Database, 〈https://ukair.defra.gov.uk/assets/ documents/reports/cat08/1212100954_31772_MPMD_Draft_Final_Report_for_ comment.pdf〉 [Accessed 20 December 2016]; 2013. [14] Suoheimo P, Grönroos J, Karvosenoja N, Petäjä J, Saarinen K, Savolahti M, et al. Päästökattodirektiiviehdotuksen ja keskisuurten polttolaitosten direktiiviehdotuksen toimeenpanon vaikutukset Suomessa, 6. Suomen Ympäristökeskuksen Raportteja; 2015, [ISBN 978-952-11-4430-1. ISSN 1796-1726. Finnish. [dataset]. [15] Data from consultations with 120 combustion plants. [16] VITO. Beste Beschikbare Technieken (BBT) voor nieuwe, kleine en middelgrote stookinstallaties, stationaire motoren en gasturbines gestookt met fossiele brandstoffen. 〈https://emis.vito.be/sites/emis.vito.be/files/pages/1142/2012/vito_BBT_Stookolie_ bookmarks-cover.pdf〉 [Accessed 20 December 2016]; 2011. [17] AEA. Assessment of the benefits and costs of the potential application of the IPPC Directive (EC/96/61) to industrial combustion installation with 20–50 MW rated thermal input. Final Report to the European Commission; 2007. [18] Amann M, Cofala J, Heyes C, Klimont Z, Mechler R, Posch M, et al. RAINS Review 2004. The RAINS model. Documentation of the model approach prepared for the RAINS peer review 2004. Austria: International Institute for Applied Systems Analysis; 2004. [19] Kelly JA. An overview of the RAINS model – environmental research centre report. Environmental Protection Agency; 2006. p. 2006. [20] Vávrová K, Knápek J, Weger J. Short-term boosting of biomass energy sources— determination of biomass potential for prevention of regional crisis situations. Renew Sustain Energy Rev 2017;67:426–36. [21] Vávrová K, Knápek J, Weger J, Králík T, Beranovský J. Model for evaluation of locally available biomass competitiveness for decentralized space heating in villages and small towns. Renew Energy 2017.
costs. The annualization was made with a 4% discount rate and a time horizon of 15 years. In our view, the obvious problem of similar impact assessment models used at the European level (such as PRIMES, GAINS, RAINS) is the high level of the automatic application of abatement measures (see [18,19]), while the case is often necessary closure or total technological change of the production process in the companies due to the otherwise certain loss of competitiveness. In such situations, abatement cost curves cannot be automatically applied but the regulatory impacts have to be assessed individually at the company level taking into respect local prices, substitutes, markets and habits. Our experience falls mostly in the field of energy and heat production, where especially the heat production is the case. In our opinion, the difference in results between the AMEC study and the Czech impact assessment is the level of detail and depth of the analysis. In our view, it is especially important to pay attention to cases where regulatory interventions are so severe that the regulated subjects lose their competitiveness. Then the regulator receives feedback that can lead to a reassessment of the initial regulatory requirements. This aspect is one of the reasons why ex-ante regulatory impact assessment is necessary. It is not only a simple calculation of costs that the regulation induces, but mainly a warning whether meeting of the regulatory requirements is feasible compared to the expected benefits. Here, it is important to have in mind the basic regulatory principle of regulatory proportionality. The benefits of the regulation do not necessarily have to be quantified in order to be able to assess adequacy of the regulation. Sometimes regulatory impact assessment (RIA) is viewed as applied cost-benefit analysis, where individual participants in the RIA process represent both the cost-bearing and benefiting bodies and where the optimal extent of regulation can be revealed at the social level. However it is necessary to have a sound analytical information for a smart regulation design, the good example in Czech Republic is the case of using biomass for energy purposes [20,21]. In our view and based on our long-term observations, the big difference between the Czech national impact assessments and the European impact estimates proves the necessity to carry out analysis at the local level. We also question the relevance of the European impact assessments where even in cases where the country provides almost all the required and needed data the impact assessment is out of reality (which in our view is the difference, tenfold). We ask ourselves if it would not be better if no assessment were made rather than pretending that the level of future impact is known. 9