Market power in a coal-based power generation sector: The case of Poland

Energy 36 (2011) 6634e6644

Contents lists available at SciVerse ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Market power in a coal-based power generation sector: The case of Poland
ski* Jacek Kamin Mineral and Energy Economy Research Institute (MEERI) of Polish Academy of Sciences, Energy and Environmental Policy Division, Wybickiego 7, 31-261 Kraków, Poland

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 April 2011 Received in revised form 1 August 2011 Accepted 28 August 2011 Available online 22 October 2011

The paper presents an analysis of market power in the Polish power sector. The study is carried out for the structure that was established by the last consolidation undertaken by the government in 2007, using a game theoretic model of the power generation market (the PolMark model). The model is run under five scenarios and eight cases. The scenarios distinguish between assumptions on strategic behaviour, whereas the cases distinguish assumptions on coal prices. The following measures are discussed in this study: electricity prices, production volumes, consumer and producer surpluses, dead weight welfare loss, CO2, SO2, NOx emissions and fuel supplies to power producers. The results confirm that the potential to exert market power in the Polish power generation sector may influence significantly electricity prices and production volumes. The analysis indicates that under the competitive scenario the average wholesale electricity price would be approximately 14.7% lower and the production would be 6.7% higher when compared to the reference scenario. Furthermore, apart from surplus transfer between producers and consumers, the dead weight loss was estimated at the level of 123.6 MV. This value reflects the net social loss resulting from uncompetitive market equilibrium.
2011 Elsevier Ltd. All rights reserved.

Keywords: Market power Electricity generation sector Market equilibrium model Cournot approach

1. Introduction Although the Polish electricity generation sector underwent several consolidations resulting in a significant change in market shares held by power companies, no quantitative analysis of market power in Poland has so far been published. Undeniably, the Energy Regulatory Office fulfils its obligation and submits national reports to the European Commission with analyses based on the use of concentration ratios and the Herfindahl-Hirschman Index (HHI).1 The results of an index-based analysis based on structural indices [1,64] indicate that the Polish utilities have the potential to exert market power. The largest electricity producer accounts for 39% of the market in terms of production volumes and 30% in terms of installed capacity, and the HHI is at the level of almost 1400 (electricity production) and 2000 (installed capacity). Also the Residual Supply Index (RSI) analysis indicates that there is a potential for market power in the power generation sector. The last significant consolidation of the state-owned part of the industry, undertaken by the government in 2007, resulted in the largest producer becoming the first pivotal supplier (RSI ¼ 0.99) [1]. However, that

conclusion per se does not give any quantitative estimation of the scale of possible abuse, and in particular does not provide insights about the impact on prices and quantities supplied. In consequence, simplified index-based tools should be applied with special care, bearing in mind their limitations. In order to correctly address the problem of market power in the power generation sector, a method that adequately captures its uniqueness and provides relevant quantitative outputs is needed. For this purpose the application of a model of the power generation market is highly recommended. The model-based method proposed above has been used in several studies. As regards the market equilibrium modelling trends, two main approaches were employed in recent years2: (i) the Cournot-based approach (also with the inclusion of Conjectural Variation) [5e14], which is well-understood, widely applied, and praised for its computational properties and tractability, (ii) the Supply Function Equilibrium (SFE) approach [15e21,61] usually more difficult in practical applications since the model is defined as a set of differential equations. As concerns the recent achievements in Cournot-based modelling of electricity markets,3 the following studies are of

2

* Tel.: þ48 12 633 02 96; fax: þ48 12 632 35 24. E-mail addresses: [email protected], [email protected]. 1 Please check [2e4] for details of the methodology of the Herfindahl-Hirschman Index, the Lerner Index, the Supply Margin Assessment (SMA), the Pivotal Supplier Indicator (PSI) and the Residual Supply Index (RSI). 0360-5442/$ e see front matter
2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2011.08.048

See [23] for a concise study on recent trends in electricity market modelling. A similar approach can be applied to other energy markets such as coal or natural gas markets (see as an example [24] for recently developed model of the international steam coal market and [29e33] for recently developed models of the natural gas market). Some remarks concerning interpretation of results of the models assuming perfectly competitive market were given in [63]. 3


ski / Energy 36 (2011) 6634e6644 J. Kamin

key importance.4 Lise et al. [22] developed a game theoretic model of the European electricity market, based on a model applied previously to the German electricity market ([7,25,26], followed by [22]). The model was extended to study the economic and environmental outcomes of various market structures in eight European countries. The variable cost of generation included fuel cost. As a result, the model was insensitive to changes in fuel prices, although by assuming various levels of costs such a feature could be introduced. An improved version of this model was employed for a study on long-term price and environmental outcomes of market reforms in the electricity market [11]. The scope of this study required an extension of time horizon which was achieved by the inclusion of a dynamic component that enabled the capture of longterm investment decisions. However, neither the first nor the second version of the model incorporated the renewable constraints of Member States which are obliged to fulfil the provisions of Directive 2001/77/EC [27] (recently amended by 2009/28/EC [28]). Market power issues in the European electricity sector, in the context of environmental constraints and transmission capacities, were also studied in [13]. For these exercises COMPETES, an EUwide model of the electricity market [8,9], was employed. In particular, the possible outcomes of dry weather, which usually leads to a decrease in hydro power production and hence an increase in electricity prices, was studied. As concerns the transmission issues, the reduction in market price which would result from a 50% increase in transmission capacity was also estimated. However, neither the amount of electricity production nor production/consumer surpluses were discussed in this paper. Linares et al. [12] proposed a generation expansion model of the power system which incorporated oligopolistic behaviour together with the CO2 emissions and the green certificates market. Since the model was applied to the Spanish electricity market special attention was given to pump storage scheduling. However, simplifications were assumed to reduce the size of the problem and assure the computability of the model. The generating units were aggregated into technology-dependent groups. Moreover, the green certificates market was not included directly within the core model, but was computed separately as a sub-model, hence the renewable electricity was subtracted from the total demand. Such an assumption might however influence the final results. Tanaka [14] carried out a market power analysis for the Japanese wholesale electricity market using a transmission-constrained Cournot model. Special attention was brought to the impact of increasing transmission capacity5 between the eastern and western part of Japan. The full cost-benefit analysis would require, however, a long-term analysis, and the focus of this paper was only on the peak period. Furthermore, the paper discussed possible outcomes of demerging the largest power producer, which would significantly reduce market power. Carraretto and Zigante [62] considered management of thermal power plants owned by strategic producers competing with other power companies and non-strategic players in the day-ahead market. They applied an optimisation technique based on the dynamic programming theory. To determine the set of optimum supply offers for each power plant an iterative procedure was employed. The model was applied inter alia to discuss the impact of market power and strategic bidding on the market equilibrium, prices and quantities.

4 A comparative analysis of the Cournot and SFE models was carried out by Willems et al. [36]. A study on the robustness of the results based on Cournot-type models was published by Neuhoff et al. [37]. 5 The technical and economic benefits from transmission capacity expansions were recently studied by Bresesti et al. [34].

6635

Discrepancies in the formulation of the aforementioned models depended mainly on the scope of the study and country-related circumstances. In fact, there is no one-fits-all methodology for developing such models and each model has to be tailored to specific conditions. Although the literature reveals that many model-based analyses have been carried out for several countries, the Polish electricity sector has not yet been the subject of any of them. The models applied were designed specifically for the countries in question, and none of them has a fuel-mix like that in Poland (more than 93% of electricity is generated by coal-fired power plants). Consequently, the emphasis is usually put on the best possible representation of electricity markets in most market equilibrium models, thus neglecting or excessively simplifying the primary energy supplies to the power generation sector (simply by including the fuel costs directly in the marginal costs). Although in specific circumstances these simplifications may be acceptable, in some cases a better representation of this element is needed, since fuel prices can fluctuate significantly, directly influencing generation costs. The inclusion of fuel price aspects in the analysis of market power issues is advisable, and of crucial importance not only for quasi mono-fuel electricity systems like the Polish one but also for others, with a more diversified fuelmix. This article also contributes to the modelling of electricity markets in this regard. Taking into consideration all these circumstances, the aims of this paper are: (i) to develop a game theory-based market equilibrium model of the Polish power generation sector (the PolMark model), and (ii) to carry out an analysis of market power in the coalbased power generation sector. The analysis presented herein is carried out for the market structure that was set up by the last consolidation undertaken by the government in 2007 (which has been maintained practically to date). In particular the impact of the strategic behaviour of power producers on electricity prices, production volumes, consumer and producer surpluses, dead weight welfare loss and fuel supplies to the power generation sector under different scenarios are considered. The remainder of the paper is organised as follows. In Section 2 a brief overview is given of the structure of the power generation sector. Section 3 presents market power in the context of the power generation sector. The methodology applied in this study is presented in Section 4. Then, the results are presented and discussed (Section 5). The paper ends with conclusions (Section 6). 2. An overview of the structure of the Polish power generation sector The total installed capacity of the Polish power generation sector amounts to 35.6 GWe (2009), which consists of public thermal power plants: 30.8 GWe, hydro power plants: 2.2 GWe, industrial power plants: 1.8 GWe and renewables: 0.8 GWe [35]. The power generation sector has undergone several consolidations and privatisations since the economic transformation started in 1989. The first important consolidation took place in 2000, when the government consolidated six State-owned hard coal-based power plants and two combined heat and power plants (CHPs) creating the Po1udniowy Koncern Energetyczny (PKE) SA (the Southern Energy Concern), at that time the largest power generation company. The total installed generation capacity was almost 5 GW (a market share of approximately 14%). The next consolidation happened in 2004 when two brown coal-based power plants: Be1chatów SA and Turów SA, were consolidated with the hard-coal fired power plant Opole SA establishing the new biggest power producer Be1chatów-OpoleTurów Górnictwo i Energetyka SA (BOT Mining and Energy). The total installed capacity of BOT amounted to 7.9 GW and its market

6636


ski / Energy 36 (2011) 6634e6644 J. Kamin

share reached 22.5%. Furthermore, two brown coal mines (KWB Be1chatów SA and KWB Turów SA) were also incorporated into BOT. Although the market share of the largest power producer (BOT) was already very high, the government again decided to consolidate the State-owned power plants. In 2006 the Government approved the principles for the energy sector policy carried out by the Ministry of Industry in the “Programme for the power sector” [38]. The key aim of this reform was to consolidate the power sector in order to ease access to the funds needed for the new investments which the Government anticipated being necessary to maintain the reliability of the power system.6 The consolidation proposed in the Programme was carried out during 2007, thus two new companies were constructed on the foundations of BOT SA and PKE SA. The biggest one, Polska Grupa Energetyczna (PGE) SA (Polish Energy Group) based on BOT, has obtained a 30% market share. The second biggest, Tauron e Polska Energia SA (Tauron Polish Energy), built upon the PKE has obtained approximately 15% of total installed capacity. Summarising, currently there are five significant electricity producers operating in the electricity market (Figs. 1 and 2), namely:  Polska Grupa Energetyczna PGE SA (Polish Energy Group), which as a result of the consolidation took control of Be1chatów Opole Turów (BOT) SA and currently has a market share of approximately 39% (in terms of electricity production) and around 30% (in the case of installed capacity);  Tauron - Polska Energia SA (Tauron - Polish Energy), which has a market share of 13.6% (electricity production) and 15% (installed capacity);  EdF Group, which actively took part in the privatisation of the Polish power plants and shares almost 10% of the market in terms of electricity production and almost 9% of installed capacity;  Enea SA, which took control of the Kozienice Power Plant after the 2007 consolidation and shares 7.5% in terms of electricity production and almost 8% of installed capacity,  Pa˛ tnów-Adamów-Konin PAK SA e which is partially privatised and shares almost 7.5% of electricity production and more than 7.5% of installed capacity.

Fig. 1. Market shares in the Polish power generation sector with respect to total installed capacity, by energy companies, 2008.

In principle, there are two components of market power that are usually addressed in the literature [46,47]: (i) physical withholding (reducing output) e when a producer reduces its offer by artificial activities, the market price turns out to be higher; (ii) financial withholding (increasing the price of a product) e when a producer increases the price of its product above the marginal cost its profits may increase, even if physical sales are reduced. In general it is very difficult to find out whether a producer applies physical or financial withholding (or both simultaneously), hence these strategies are sometimes assumed to be equivalent [46]. The exertion of market power not only increases the producer surplus, but also decreases the consumer surplus. Unfortunately this is not the end of the story, since apart from the transfer of surplus, the total net social surplus decreases and a dead weight welfare loss appears. In fact, most papers discuss the impacts on prices and quantities, but the welfare analysis is neglected. However, such an analysis is of crucial importance, in particular when possible mergers are studied by the regulatory authorities. The outcomes of the welfare analysis could be compared to the declared synergy savings, which are usually proclaimed by the representatives of companies to be merged.

3. Market power in the power generation sector 4. Methodology Although it is possible to encounter a great number of definitions of market power in the literature,7 the most widely accepted one seems to be the one by Mas-Colell et al. [41], who define the market power as “the ability to profitably alter prices away from competitive levels”. While there have been discussions on whether this ability has to be exerted on a regular basis or only at a specific point in time, taking into consideration the particular features of power markets (characterised by inelastic demand, limited availability of efficient storage technologies and significant changes in demand at different hours/seasons) one is led to the conclusion that it is perfectly sufficient for a company behaving strategically to only exert market power for a short period when the conditions are most favourable (usually at peak hours). As a result the company increases its producer surplus causing a reduction in consumer surplus (the Californian case is a good example).

6 Unfortunately, the Programme [38] contained contradictory provisions in some parts. As an example: one of the outcomes resulting from its implementation (more consolidation) was supposed to be an increase in competition in the electricity market. 7 The theory of market power and its application to power systems has been profoundly discussed in the literature, and the number of references is enormous (for example: [39e49]). As a result, to omit extensive repetitions, only the key aspects of market power are presented in Section 3.

Game theory-based market equilibrium models of the electricity generation sector have, as already pointed out, been developed for several national and international markets. However, such a model has never been developed for Poland. Taking into consideration the advantages and disadvantages of the most common methodologies,

Fig. 2. Market shares in the Polish power generation sector with respect to gross electricity production, by energy companies, 2008.


ski / Energy 36 (2011) 6634e6644 J. Kamin

the Cournot approach, improved by the inclusion of Conjectural Variations, was chosen for this study. Classic Cournot models were severely criticised in the past, because, as a result of neglecting competitors’ supply functions, the prices were too high and the production volumes were too low when compared to actual market values. An improvement was therefore needed with regard to the standard formulation of the Cournot model. The Conjectural Variation for a given power producer is defined as its expectation of the competitors’ reaction to its production decision. Enriching the Cournot approach with the inclusion of Conjectural Variations increases the plausibility of the model outcomes and leads to better insights into the problem. Cournot-based models are defined by a set of algebraic equations, which makes them more tractable and eases considerably the resolution of the problem. A Supply Function Equilibrium (SFE) model, on the contrary, is defined as a set of differential equations which are much more troublesome to deal with and solve (difficulties in calibration, many equilibriums). That difference is of great importance when geographically extensive or richly detailed markets are modelled. Conjectural Variations (CV) also ease the process of defining the different strategies of the power producers. By assuming different values of the CV factors the market participants can be assumed to behave competitively (as price-takers) or strategically (fully, according to the Cournot assumption, or with lower intensity). Also, several market structures can be reflected in a model, such as perfect (price) competition, oligopoly (strategic behaviour) or a cartel. In fact, the equilibrium in the real markets is usually between the competitive (Bertrand) and the full strategic (Cournot) equilibrium. Summarising, the Cournot approach with the inclusion of Conjectural Variation forms a flexible and powerful methodology that can be applied to develop realistic and useful models of electricity markets. The market equilibrium game theory-based model which is employed for this study (the PolMark model) reflects the most important relations existing in the Polish power market. Owing to the aim of the study, the generation sector is represented at power plant level. Power plants operate as individual units consolidated within energy groups and holdings. The remaining power plants operate on an individual basis. In detail, there are 14 hard coalbased public power plants, 6 brown coal-based public power plants and 33 public CHP (Combined Heat and Power) plants represented in the model. Due to their limited market penetration, renewable-based power plants are aggregated in technologydependent groups. As concerns the demand for electricity, four seasons (winter, spring, summer and autumn) and three loads (base, mid, peak) are included in this version of the model (resulting in 12 different load types). The model captures different market structures dependent on the strategies of the market players. As far as fuel supplies are concerned, fuel prices are assumed to be subject to quarterly fluctuations, hence leading to a more precise representation of the fuel prices / electricity generation costs / electricity prices relationship. By including this, the model enables analyses to be made of the impact of the different behaviour of power generation companies with regard to fuel supplies to the power sector. This model is fully capable of supporting research focused on renewables, and in particular on the performance and outcomes of the green certificates market. Conversely to other models of this type the green certificate market is incorporated directly within the model (not as a sub-model). 4.1. Mathematic formulae of the model Following microeconomic theory ([39,42e44]) each power generation company sets its production level at the point where the first order condition (the first derivative of the profit function with respect to quantity produced) equals zero, which is mathematically

6637

expressed as: vPf =vqf ¼ 0. For the Cournot-based model with inclusion of Conjectural Variations this condition takes the following form ([50,51]):

 vPf qf q,vp
vcf 1 þ CVf  ¼ p þ p, , vqf q p,vq vqf   sf
vcf 1 þ CVf  ¼ p, 1 þ ¼ 0 3 vqf where: f e index of electricity producer, p e market price for electricity, qf e electricity production, CVf e conjectural variation, sf e market share, vcf =vqf e marginal costs, 3 e price elasticity of demand. Based on the expression presented above, the mathematical representation of the PolMark model is hereafter given in the following form: first the indexes, parameters and variables are listed (Table 1), then the power market equilibrium conditions are

Table 1 List of indexes, parameters and variables used in the model formulae. Indexes p o c m l

f q PF CP LQ Parameters Capacity Factorp,l CVc,p Durationl Efficiencyp Elasticityl Emission Factorm,p Emission Limitm Fuel Pricef,q Installed Capacityp Lossesp MinRenShare MSharec,p Reference Demandl Reference Pricel Renewable Indicatorp

VariableCostp,l Variables Pricel Productionp,l Profitp Shp Emissionm Shp Capacityp,l Shp Renewables

Power plants and CHPs, p ˛ P Renewable power plants and CHPs, o ˛ P Energy groups, c ˛ C Pollutants, m ˛ M ¼ {SO2, PM, NOX, CO2} Load, l ˛ L ¼ {WinterPeak, WinterMid, WinterBase, SpringPeak, SpringMid, SpringBase, SummerPeak, SummerMid, SummerBase, AutumnPeak, AutumnMid, AutumnBase} Fuel, f ˛ F ¼ {HC, BC, GAS, OIL,.} Quarter, q ˛ Q ¼ {Winter, Spring, Summer, Autumn} Assignment of fuels f which are permitted to use in power plant p Assignment of power plant p to energy group c Assignment of quarters q to loads l Capacity factor of power plant p in load l [%] Conjectural variation factor of power plant p controlled by energy group c Duration of load l [h] Efficiency of electricity production by power plant p [%] Load dependent price elasticity of demand [%] Emission factor for pollutant m in power plant p [kt/MWh] ([Mt/MWh] for CO2) Annual emission limit for pollutant m [kt] ([Mt] for CO2) Price of fuel f in quarter q [z1/MWh] Installed capacity of power plant p [MW] Losses of power plant p [%] Minimal share of renewable electricity production [%] Market share of energy group c that controls power plant p [%] Reference demand for electricity in load l (historical data for the year in question) [MW] Reference market price for load l (historical data for the year in question) [MW] Indicates if power plant p is a renewable-based one (1 e renewable power plant, 0 e non-renewable power plant) Non-fuel related variable cost of power production in power plant p in load l [z1/MWh] Market price of electricity in load l [z1/MWh] Electricity production of power plant p in load l [MW] Profit of power plant p [z1] Shadow price of emission constraint [z1/kt of pollutant m] ([z1/Mt of CO2]) Shadow price of capacity constraint in load l in power plant p [z1/MWh] Shadow price of renewable electricity production constraint [z1/MWh]


ski / Energy 36 (2011) 6634e6644 J. Kamin

6638

presented with the corresponding Lagrange function. Finally, the list of Karush-Kuhn-Tucker (KKT) conditions (the set of equations defining the model) is given. Each power plant maximises its profit by choosing the level of power production expressed in [MW] according to the following profit function:

Profitp ¼

P l



  Durationl ,Productionp;l , 1  Lossesp ,Pricel þ

P

0

Durationl ,Productionp;l

l

,@VariableCostp;l þ

1 X FuelPricef ;q A Efficiencyp q:q˛LQ

X f :f ˛PF

Lp ¼ 

P

  Durationl ,Productionp;l , 1  Lossesp ,Pricel þ

l

þ

P l

þ

P

Durationl ,Productionp;l , VariableCostp;l þ

P f :f ˛PF

X

Durationl ,

X

Productiono;l  MinRenShare

o˛P

l

,

X

Productionp;l

Following the methodology of development of market equilibrium models, the maximisation problem was transformed into a minimisation problem before deriving the KKT conditions. As a result, the optimisation problem is represented by the following Lagrangian (Lp):

1 P FuelPricef ;q Aþ q:q˛LQ Efficiencyp


 Durationl  ShpCapacityp;l , Productionp;l  InstalledCapacityp;l ,CapacityFactorp;l þ

þ þ

m

P

ShpEmissionm ,

P

Durationl ,

P

l

P

Durationl ,ShpRenewables,

p

l

p

Productionp;l ,MinRenShare 

  Productionp;l , 1  Lossesp ¼ ReferenceDemandl

p



Pricel ReferencePricel

,

X vLp ¼ VariableCostp;l þ vProductionp;l

f :f ˛PF

þShpCapacityp;l þ

P m

!

Productionp;l ,EmissionFactorm;p  EmissionLimitm þ

The demand is price-elastic, hence the demand function takes the following form:

X

0

p

l

P

!

Elasticityl

P p

! Productionp;l ,RenewableIndicatorp

The optimal Karush-Kuhn-Tucker (KKT) conditions are derived by taking the first derivative with respect to the production variable, namely: Productionp,l. To incorporate the impact of market power on price-cost relations (mark-up) in the model, the conjectural variation factor (CV) was introduced. The list of KKT conditions reads, therefore, as follows:  Derivative of the Lagrangian (Lp) e the profit function:

X FuelPricef ;q þ Efficiencyp q:q˛LQ

ShpEmissionm ,EmissionFactorp;m  ShpRenewables,RenewableIndicatorp þ

     MSharec;p  , 1 þ CVc;p  0t0  Productionp;l Pricel , 1  Lossesp , 1 þ Elasticityl

The capacity constraint is defined as follows:

where MSharec,p is calculated on the basis of installed capacity for the year analysed, expressed as:

P

InstalledCapacityp;l ,CapacityFactorp;l  Productionp;l  0

p:p˛C

The emission constraint is defined for each pollutant m separately, and reads as follows:

EmissionLimitm 

X l

Durationl ,

X

MSharec;p ¼ P p

InstalledCapacityp

InstalledCapacityp

 Power demand function:

Productionp;l

p

,EmissionFactorm;p  0 Since the Polish power sector is obliged to fulfil the obligation to produce a certain amount of renewable electricity, the following constraint is needed:

X p

  Productionp;l , 1  Lossesp

Elasticityl  Pricel  ReferenceDemandl , ReferencePricel ¼ 0;

Pricel free


ski / Energy 36 (2011) 6634e6644 J. Kamin

 Capacity constraint:

Table 2 Scenario and case acronyms.

InstalledCapacityp;l ,CapacityFactorp;l  Productionp;l  0t0

Scenario/case

 ShpCapacityp;l

Scenario: REF

 Emission constraint:

COMP FULL1ST

 Renewable electricity production constraint:

P

Case (formed by adding the following suffix): _HC20UP Hard coal prices increased by 20% _HC20DN Hard coal prices decrease by 20% _BC20UP Brown coal prices increased by 20% _BC20DN Brown coal prices decrease by 20%

! Productionp;l ,Durationl ,RenewableIndicatorp

p;l

P l

P 

ALLSAME

p

 0t0  ShpEmissionm

Description Reference scenario based on market outcomes for the year in question Competitive behaviour of all electricity producers Full strategic behaviour of the biggest electricity producer (all other producers behave as price takers) Full strategic behaviour of the 2nd biggest electricity producer (all other producers behave as price takers) All electricity producers are assumed to behave strategically with the intensity of the biggest producer under the REF scenario

FULL2ND

EmissionLimitm X X  Durationl $ Productionp;l $EmissionFactorm;p l

6639

Durationl !

Productionp;l ,Durationl ,MinRenShare

p;l

P

Durationl

l

 0t0  ShpRenewables

4.2. The scenarios Five scenarios and eight cases are assumed in this study. The scenarios distinguish between assumptions on strategic behaviour, whereas the cases distinguish assumptions on coal prices. The REF scenario is based on historical data for the year in question (2008). It assumes strategic behaviour of the biggest market player with its intensity reflected by the CV factor. The other power producers are assumed to be price takers. The COMP scenario presumes no market power potential in the electricity market, all the power producers behave competitively and the market price is set purely on the cost basis. The FULL1ST scenario assumes full strategic behaviour of the biggest power producer which exerts in full its market power (under the Cournot assumption). The other players behave as price takers. Similarly, the FULL2ND scenario assumes the second biggest power producer behaves strategically, fully exerting its market power, and the remaining generation companies behave as price takers. Furthermore, to study the impact of hard coal prices, which are important for the power plants, cases assuming changes in the fuel prices were introduced. The _HC20UP, _HC_20DN, _BC20UP and _BC20DN cases assume a 20% change in the price of hard (HC) and brown coal (BC). The differentiation between hard coal and brown coal cases is needed because brown coal is supplied to the power producers solely from mines located very close to the power stations (and its price is mostly based on the cost of production) while hard coal is supplied from domestic mines and is also imported (and prices are dependent on developments in the world coal market). The cases are run only for the REF and the COMP scenarios. The full list of scenarios and cases is presented in Table 2. 4.3. Main data assumptions Since the scope of data used in this model is broad, only the most important data are presented below. As previously stated, the

power generation sector is represented in the model at power plant level, although in the following tables the aggregated data are presented. The techno-economic parameters such as installed capacity, efficiency and non-fuel related variable costs, are listed in Table 3, and the emission factors in Table 4. The average season-dependent load curves are derived from the Polish transmission system operator (PSE-Operator SA) data. The load profile is constructed for the wholesale electricity market in the following way. The yearly load curve (8760 hours) is split into seasons (quarters). Then separately for each quarter the load duration curve is constructed, which is subsequently split in a way that allows to capture peak, mid and base loads. In consequence, there are 12 loads incorporated in the model represented by demand for power and load duration (Table 5). The estimation of the Conjectural Variation (CVi) factors [54,55] is based on historical data. The CVi is estimated on the basis of the following relationship:

Li ¼

si ð1 þ CVi Þ 3

¼ >CVi ¼

3 ,Li

si

1

Fuel sector data are mostly collected from [52] and [56,57]. 4.4. Implementation of the model The model is developed in GAMS (General Algebraic Modelling System) [58] as a Mixed Complementary Problem (MCP) and the solution is found by the PATH solver [59,60]. The model is calibrated to the Polish power generation sector and the wholesale electricity market. The year in question is 2008, since this is the first full year of operation of all the newly established energy companies after the consolidation that took place during 2007. The monetary values in the model are expressed in the national currency (z12008),

Table 3 Aggregated techno-economic parameters of the fossil fuel-based public power generation sector for 2008. Power plant group

Installed capacity [MW]

Gross efficiency [%]

Non-fuel related variable costs [V/MWh]

Hard coal public power plants Brown coal public power plants

15607.0

37.9

1.7

9280.0

37.6

4.0

Source: [35,52].


ski / Energy 36 (2011) 6634e6644 J. Kamin

6640

Table 4 Average emission factors of the fossil fuel-based public power generation sector, kg/MWh (t/MWh for CO2) 2008. Power plant group

SO2

NOx

CO2

Hard coal public power plants Brown coal public power plants

2.1819 3.7202

1.6022 1.3270

0.8823 1.0599

Table 6 Electricity production [TWh] and wholesale price of electricity [V/MWh] under the REF, COMP, FULL1ST, FULL2ND and ALLSAME scenarios, 2008. REF

COMP FULL1ST FULL2ND ALLSAME

Electricity production [TWh] 147.97 157.83 108.61 Average market price [V/MWh] 44.1 37.6 94.3

143.06 47.7

140.73 49.6

Source: [35,53].

and all results presented in this paper are expressed in V2008. A series of sensitivity analyses were carried out to validate the behaviour and robustness of the model. The results of the sensitivity analysis confirmed that the model correctly responds to changes imposed on the input data. 5. Results and discussion Several measures have been applied to carry out the analysis of market power in the power generation sector. In particular the following measures are discussed under the scenarios and cases assumed beforehand: (1) gross electricity production [TWh], (2) market price of electricity [V/MWh], (3) consumer surplus [MV], (4) producer surplus [MV], (5) net social surplus [MV], (6) dead weight welfare loss [MV], (7) electricity production fuel-mix [TWh], (8) SO2, NOx, and CO2 emissions from the power generation sector [kt] ([Mt] for CO2) and (9) fuel supplies to the power sector (separately for hard coal, brown coal, etc.) [PJ]. The reference scenario (REF) assumes strategic behaviour by the largest player, with the intensity derived from the market outcomes of 2008. Under this scenario the total electricity production equals 147.97 TWh and the market price is 44.1 V/MWh. In a fully competitive electricity market the price of electricity is based on the cost of the marginal supply, and no market power mark-up is applied by any producer (the COMP scenario). A comparison of such a scenario with the reference one (REF) shows that electricity consumption would increase, leading to an increase in total electricity production by 9.86 TWh. This would be accompanied by a decrease in electricity price to 37.6 V/MWh (6.5 V/MWh lower than in the REF scenario) (Table 6). Under the scenario assuming full strategic behaviour of the largest power producer, according to the Cournot assumption (the FULL1ST scenario), the electricity price would soar to 94.3 V/MWh suppressing electricity consumption. As a result the total production would plummet to only 108.61 TWh. Such a scenario would have a significant impact on the surplus transfer, to be discussed later on. The scenario under which the 2nd biggest power producer behaves fully strategically would cause a drop in electricity production by only 4.91 TWh and an increase in electricity price by 3.6 V/MWh. Interestingly, if all the power producers are assumed to behave strategically with the same intensity as the biggest power producer in 2008 (the ALLSAME scenario), the electricity production would decrease by 7.24 TWh causing an increase in price by 5.5 V/MWh (Table 6).

Table 7 Consumer, producer, net social surplus and dead weight loss under the REF, COMP, FULL1ST, FULL2ND and ALLSAME scenarios [MV], 2008. REF

Difference when compared to the REF scenario: COMP

Consumer Surplus Producer Surplus

244 357.7 3 430.7

Net social surplus

247 788.3

FULL1ST

913.8 -5 772.8 790.2 4 532.7 Dead Weight Loss 123.6 -1 240.1

FULL2ND

ALLSAME

468.5 434.5

710.2 655.2

34.0

55.0

Under the REF scenario the consumer and producer surpluses are estimated to be at the level of 244.4 and 3.4 billion V respectively. In consequence, the total net surplus amounts to 247.8 billion V. If all the market players had behaved competitively (the COMP scenario) the consumer surplus would have increased by 0.91 billion V and the producer surplus would decrease by 0.79 billion V. Having calculated the producer and consumer surpluses, an estimate of the dead weight welfare loss is possible under scenarios assuming various behaviours of power producers. The dead weight loss of the REF scenario when compared to the COMP scenario is estimated to be at the level of 123.56 MV. This reflects the net social loss resulting from market power in the power generation sector (Table 7). Full strategic behaviour by the largest power producer (FULL1ST) would result in an immense surplus transfer from consumers to producers, causing a dead weight welfare loss increase of 1.24 billion V when compared to the REF scenario. Strategic behaviour of the second biggest market player (FULL2ND) would not be that severe for the consumers, as it would lead to a decrease in their surplus by 0.45 billion V and an increase in dead weight loss by almost 34.0 MV. Under a scenario in which all producers take the same strategy trying to apply the mark-up at the same level (ALLSAME), the reduction of consumer surplus of 0.71 billion V would be accompanied by an increase in producer surplus of 0.65 billion V, with an increase in dead weight welfare loss of 54.9 MV (Table 7). As the power producers operating in the Polish power generation sector are characterised by diverse generation fuel-mixes (mostly related to the hard/brown coal split), scenarios assuming changes in market behaviour also influence the overall fuelgeneration mix. As already mentioned, under the COMP scenario the overall electricity production would increase. The total increase by 9.86 TWh would be almost equally distributed between hard coal and brown coal-based electricity production (4.22 and

Table 5 Average hourly season-dependent load and load duration of the Polish power system, 2008. Load

Base Mid Peak Total

Winter

Spring

Summer

Autumn

Average load

Load duration

Average load

Load duration

Average load

Load duration

Average load

Load duration

[MW]

[h]

[MW]

[h]

[MW]

[h]

[MW]

[h]

14765.9 19659.7 23965.7 58391.3

365 1787 32 2183.0

13114.2 17267.9 20924.2 51306.3

364 1783 37 2184.0

13111.9 17252.8 21177.0 51541.6

421 1749 38 2208.0

14222.9 19106.7 23231.5 56561.0

482 1699 28 2209.0

Source: own calculations based on data derived from PSE Operator SA.


ski / Energy 36 (2011) 6634e6644 J. Kamin

6641

Table 8 Electricity production fuel-mix of the Polish power sector under the REF, COMP, FULL1ST, FULL2ND and ALLSAME [TWh], 2008.

Total electricity production, of which: Hard coal Brown coal Natural gas Renewables

REF

COMP

FULL1ST

FULL2ND

ALLSAME

147.97

157.83

108.61

143.06

140.73

87.40 54.42 2.47 3.69

91.62 58.77 3.68 3.77

71.27 31.64 1.99 3.70

80.11 55.45 3.82 3.69

79.16 55.45 2.42 3.70

Table 9 Nominal change in electricity production fuel-mix in comparison with the REF scenario [TWh], 2008. REF COMP FULL1ST FULL2ND ALLSAME Total electricity production, of which: e Hard coal e Brown coal e Natural gas e Renewables e

9.86 4.22 4.36 1.21 0.08

39.36 16.13 22.77 0.48 0.02

4.91 7.29 1.03 1.35 0.00

7.24 8.24 1.03 0.05 0.02

4.36 TWh respectively) while generation based on natural gas would only increase by 1.2 TWh. Fully strategic behaviour of the largest power producer (the FULL1ST scenario) would result in a significant decrease in brown coal electricity production (by almost 22.8 TWh), and hard coal-based production (by 16.1 TWh). On the other hand, under the scenario assuming the second biggest power producer behaving strategically (FULL2ND) a reduction of 7.3 TWh in the case of hard coal and an increase of 1 TWh in brown coal power plants would be observed. The change in fuel-mix for electricity generation would be very similar to the one assuming the same level of strategic behaviour by all market players (the ALLSAME scenario), as a decrease in hard coal-based (by 8.2 TWh) would be accompanied by an increase of approx. 1 TWh of brown coal-based production. However, almost no change in natural gasbased electricity production would be noticed (Table 8 and 9). In principle, emissions from the power generation sector reflect the changes in electricity production. An increased production under the competitive (COMP) scenario would lead to an increase in SO2 and NOx emissions to 471.8 kt and 237.7 kt respectively. On the contrary, for scenarios where a decrease in production is observed, a reduction in emission levels is also noticed. In particular, under the FULL1ST scenario, a significant fall in emissions is observed (to 318.1 kt in the case of SO2 and to 166.3 kt in the case of NOx) (Fig. 3) as a result of reduced production.

Fig. 3. SO2 and NOx emissions under the REF, COMP, FULL1ST, FULL2ND and ALLSAME scenarios, [t], 2008.

Fig. 4. CO2 emissions under the REF, COMP, FULL1ST, FULL2ND and ALLSAME scenarios, [Mt], 2008.

A similar pattern is observed in the case of CO2 emissions. Increased electricity consumption under the competitive (COMP) scenario leads to a CO2 emissions increase of 9.8 Mt. The remaining scenarios show a decrease in CO2 emissions, with the FULL1ST scenario being the most severe (by 38.8 Mt). It is worth adding that such a “gain” in CO2 emissions reduction would be accompanied by a huge transfer of surplus from consumers to producers and a significant increase in dead weight welfare loss (Fig. 4). To discuss the impact of fuel prices on CO2 emissions under different market structures a number of cases were run. In fact, the cases in which a 20% increase in fuel price is assumed (namely the _HC20UP and the _BC20UP) could be also considered as the ones assuming a carbon tax imposed on hard or brown coal. In principle, the level of CO2 emissions under the REF scenarios seem to be insensitive to a change in hard coal price since only a change in emissions of approx. 0.5 Mt of CO2 is noticed. In the case of an increase in brown coal price, a decrease in CO2 emissions by 2.3 Mt is observed (Fig. 5). Interesting insights are derived from the analysis of the COMP cases. A 20% change in brown coal price (both up and down) does not lead to any change in CO2 emissions. This happens because, under the conditions appertaining in 2008, brown coal-based power production was the most cost-efficient option for power generation. As a result, a change in brown coal price under the COMP cases (although reasonably assumed) is not enough to cause a change in CO2 emissions. On the other hand, an increase in hard coal price by 20% would cause a drop in emissions by 2.6 Mt of CO2

Fig. 5. Change in CO2 emissions when compared to the REF scenario, [Mt], 2008.


ski / Energy 36 (2011) 6634e6644 J. Kamin

6642

Fig. 6. Change in CO2 emissions when compared to the COMP scenario, [Mt], 2008.

Table 10 Fuel supplies to the public power generation sector (for electricity production), [PJ], 2008.

Hard Coal Brown Coal Natural Gas Renewables Total

REF

COMP

FULL1ST

FULL2ND

ALLSAME

755.35 516.71 19.11 17.86 1309.04

790.34 558.04 28.45 17.54 1394.37

604.09 300.96 15.43 18.06 938.55

685.95 526.45 29.57 17.86 1259.84

676.62 526.45 18.72 18.06 1239.85

while a decrease in hard coal price by 20% would only increase the emissions by 1.6 Mt (Fig. 6). These numbers seem to be insignificant when compared to the total CO2 emissions from the electricity generation sector of approx. 142 Mt. However, it is important to keep in mind that the Polish power generation sector still remains virtually completely coal-based, and even the imposition of a carbon tax would not lead to a significant change in CO2 emissions in the short term. As previously noted, coal retains a crucial role as the main fuel for electricity generation, and the coal industry still supplies significant amounts of fuel to the power sector. As a result, the scale of demand for primary energy is of the utmost importance for the economy. Reasonably, an increase in electricity consumption observed under the COMP scenario results in an increased demand for primary fuels, by 85.3 PJ. The increase in demand for hard coal, brown coal and natural gas is by 35.0, 41.3 and 9.3 PJ respectively. Full strategic behaviour of the largest power producer (the FULL1ST scenario) would cut the fuel supplies considerably, to the level of

938.6 PJ. Withholding of brown coal-based production would reduce the demand for brown coal to approx. 300 PJ. Additionally, a reduction of approx. 150 PJ of hard coal would be observed. Fully strategic behaviour of the second biggest power producer (the FULL2ND scenario) would lead to an increase in demand for brown coal to 526.5 PJ. At the same time the total consumption of primary fuels would be approx. 50 PJ lower (Table 10). In order to capture the impact of fuel prices on the fuel supplies to the power sector under various market structures, the cases differentiating fuel prices were analysed. Under the REF scenario, hard coal supplies change insignificantly when brown coal price is changed. On the contrary, an increase in hard coal price by 20% (as already discussed this increase in price can be also treated as a carbon tax imposed on coal) leads to an 8 PJ decrease in hard coal supplies. A similar reaction is observed when the impact of the brown coal price on its supplies is concerned. The brown coal supplies are unresponsive to a change in hard coal price, but react considerably when a 20% increase in brown coal price is assumed (a reduction by 22.7 PJ) (Fig. 7). 6. Conclusions The results presented in this paper confirm that the existing potential to exert market power in the Polish power generation sector may have a significant impact on wholesale electricity prices and production volumes. The analysis indicates that under the competitive scenario the average wholesale electricity price would be approximately 6.5 V/MWh lower when compared to the reference scenario. As a result of increased consumption, the total electricity production would be higher, by almost 10 TWh. In consequence, the consumer surplus would be 0.91 billion V lower and the producer surplus would increase by 0.79 billion V. Such a surplus transfer would be accompanied by the dead weight welfare loss, which is estimated at the level of 123.6 MV. This amount reflects the net social loss resulting from market power in the power generation sector. A power generation industry solely based on coal significantly reduces the responsiveness of the sector to changes in fuel prices. The cases assuming various levels of both hard and brown coal prices confirmed that the electricity production is almost insensitive to a change in fuel prices in the short term. The effect of imposing a carbon tax (which under this modelling approach is equivalent to an increase in fuel prices) would not lead to a significant change in CO2 emissions in the short-run. The analysis of the impact of CO2 emissions on the change of fuel-mix under scenarios assuming different levels of market power would, however, require a long-term model.

Fig. 7. Hard coal and brown coal supplies to the electricity generation sector under REF, REF_BC20DN, REF_BC20UP, REF_HC20DN and REF_HC20UP scenarios, [PJ], 2008.


ski / Energy 36 (2011) 6634e6644 J. Kamin

Electricity markets, and mergers or consolidations of power generation companies in particular, shall stay under supervision of the public authorities (regulators, competition offices and appropriate ministries) in order to avoid malfunction of the markets or undesired and unexpected outcomes resulting from market reforms. Market failures or inefficient market mechanisms have a direct impact on the final consumers perceived not only via directly recognizable prices and quantities, but also via the surplus transfers and dead weight loss. Consequently, adequate policy actions seem to be necessary to protect final consumers and smaller producers against the potential oligopolistic behaviour of the biggest energy companies. Acknowledgements I would like to thank Ignacio Hidalgo González and anonymous referees for their thoughtful suggestions and comments on this paper. References
ski J. Konsolidacja sektora wytwarzania energii elektrycznej w Polsce: [1] Kamin wyniki analizy wska
znikowej. Polityka Energetyczna, Tom 2011;14. zeszyt 1, IGSMiE PAN, Kraków 2011. [2] Shukla UK, Thampy A. Analysis of competition and market power in the wholesale electricity market in India. Energy Policy May 2011;39(5): 2699e710. [3] Asgari MH, Monsef H. Market power analysis for the Iranian electricity market. Energy Policy October 2010;38(10):5582e99. [4] Chang Y. The new electricity market of Singapore: regulatory framework, market power and competition. Energy Policy January 2007;35(1):403e12. [5] Yang Ch W, Hwang MJ, Sohng SN. The Cournot competition in the spatial equilibrium model. Energy Economics March 2002;24(2):139e54. [6] Day CJ, Hobbs BF, Pang J. Oligopolistic competition in power networks: a conjectured supply function approach. IEEE Transactions on Power Systems 2002;17(3):597e607. [7] Lise W, Kemfert C, Tol RSJ. The German electricity marketddoes liberalisation bring competition? Milan, Italy: Nota di Lavoro; 2003. [8] Hobbs BF, Rijkers FAM. Strategic generation with conjectured transmission price responses in a mixed transmission pricing system e part I: formulation. IEEE Transactions on Power Systems May 2004;19:707e17. [9] Hobbs BF, Rijkers FAM, Wals AF. Strategic generation with conjectured transmission price responses in a mixed transmission pricing systemdpart II: application. IEEE Transactions on Power Systems May 2004;19(2). [10] Liu Y, Ni YX, Wu FF, Cai B. Existence and uniqueness of consistent conjectural variation equilibrium in electricity markets. International Journal of Electrical Power & Energy Systems July 2007;29(6):455e61. [11] Lise W, Kruseman G. Long-term price and environmental effects in a liberalised electricity market. Energy Economics March 2008;30(2):230e48. [12] Lineares P, Santos FJ, Ventosa M, Lapiedra L. Incorporating oligopoly, CO2 emissions trading and green certificates into a power generation expansion model. Automatica 2008;44:1608e20. [13] Lise W, Hobbs BF, Hers S. Market power in the European electricity market e The impact of dry weather and additional transmission capacity. Energy Policy 2008;36:1331e43. [14] Tanaka M. Transmission-constrained oligopoly in the Japanese electricity market. Energy Economics 2009;31:690e701. [15] Klemperer PD, Meyer MA. Supply function equilibria in oligopoly under uncertainty. Econometrica 1989;57(6):1243e77. [16] Baldick R, Grant R, Kahn E. Theory and application of linear supply function equilibrium in electricity markets. Journal of Regulatory Economics March 2004;25(2):143e67. [17] Rudkevich A. On the supply function equilibrium and its application in electricity markets. Decision Support Systems 2005;40. [18] Morris JR, Oska D. The likely effect of the proposed Exelon-PSEG Merger on wholesale electricity prices. The Electricity Journal JanuaryeFebruary 2008; 21(1):45e54. [19] Holmberg P. Numerical calculation of an asymmetric supply function equilibrium with capacity constraints. European Journal of Operational Research 16 November 2009;199(1):285e95. [20] Gao F, Sheble GB. Electricity market equilibrium model with resource constraint and transmission congestion. Electric Power Systems Research January 2010;80(1):9e18. [21] Ciarreta A, Espinosa MP. Supply function competition in the Spanish wholesale electricity market. Energy Journal 2010;31(4). IAEE. [22] Lise W, Linderhof V, Kuik O, Kemfert C, Ostling R, Heinzow T. A game theoretic model of the Norhwestern European electricity market e market power and the environmental. Energy Policy 2006;34:2123e36.

6643

[23] Ventosa M, Baíllo A, Ramos A, Rivier M. Electricity market modelling trends. Energy Policy 2005;33:897e913. [24] Haftendorn C, Holz F. Modeling and analysis of the international steam coal trade. The Energy Journal 2010;31(4). [25] Kemfert C, Tol RSJ. The liberalisation of the German electricity marketdmodelling an oligopolistic structure by a computational game theoretic model; 2000. Working Paper 89/2000, Oldenburg. [26] Kemfert C, Barbu D, Kalshnikov V. Economic impact assessment of the European electricity market; 2002. Working Paper, Oldenburg. [27] Directive 2001/77/EC of the European Parliament and of the Council of 27 September 2001 on the promotion of electricity produced from renewable energy sources in the internal electricity market. [28] Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. [29] Egging R, Holz F, Gabriel SA. The World Gas Model: a multi-period mixed complementarity model for the global natural gas market. Energy October 2010;35(10):4016e29. [30] Egging R, Gabriel SA, Holz F, Zhuang J. A complementarity model for the European natural gas market. Energy Policy July 2008;36(7):2385e414. [31] Holz F, von Hirschhausen Ch, Kemfert C. A strategic model of European gas supply (GASMOD). Energy Economics 2008;30:766e88. [32] Lise W, Hobbs BJ. Future evolution of the liberalised European gas market: simulation results with a dynamic model. Energy July 2008;33(7): 989e1004. [33] Egging R, Gabriel SA. Examining market power in the European natural gas market. Energy Policy November 2006;34(17):2762e78. [34] Bresesti P, Calisti R, Cazzol MV, Gatti A, Provenzano D, Vaiani A, et al. The benefits of transmission expansions in the competitive electricity markets. Energy March 2009;34(3):274e80. [35] Polish energy statistics 2008. Warsaw: Energy Market Agency; 2009. [36] Willems B, Rumiantseva I, Weigt H. Cournot versus supply functions: what does the data tell us? Energy Economics 2009;31:38e47. [37] Neuhoff K, Barquin J, Boots MG, Ehrenmann A, Hobbs BF, Rijkers FAM, et al. Network-constrained Cournot models of liberalized electricity markets: the devil is in the details. Energy Economics May 2005;27(3):495e525. [38] Programme for the power sector, 2006. Warsaw: Ministry of Industry; 2006. [39] Estrin S, Laidler D, Dietrich M. Microeconomics. 5th ed. Pearson Education Limited; 2008. [40] Neumann M, Weigand J, editors. The international handbook of competition. Cheltenham, Northampton: Edward Elgar; 2004. [41] Mas-Colell A, Whinston MD, Green JR. Microeconomic theory. New York: Oxford University Press; 1995. [42] Hardwick P, Khan B, Langmead J. An introduction to modern economics. 4th ed. New York: Longman London; 1994. [43] Tirole J. The theory of industrial organization. Cambridge, Massachusetts e London, England: The MIT Press; 1993. [44] Koutsoyiannis A. Modern microeconomics. 2nd revised ed. Palgrave Macmillan; 1979. [45] Helman U. Market power monitoring and mitigation in the US wholesale power markets. Energy MayeJune 2006;31(6e7):877e904. [46] Stoft S. Power system economics, designing markets for electricity. WileyInterscience: IEEE Press; 2002. [47] Hunt S. Making competition work in electricity. New York: John Wiley & Sons, Inc; 2002. [48] Rothwell G, Gomez T. Electricity economics e regulation and deregulation. New York: IEEE Press, John Wiley & Sons, Inc; 2003. [49] Belyaev LS. Electricity market reforms e economics and policy challenges. Springer; 2011. [50] Varian HR. Microeconomic analysis. 3rd ed. W. W. Norton & Company, Inc; 1992.
ski J. Równowaga rynków energii elektrycznej i paliw: podej
scie [51] Kamin modelowe. Rynek Energii nr 2011;1(92). Kaprint. [52] Techno-economic situation in the power sector. Warsaw: Energy Market Agency; 2008. [53] Emitor 2008. Warsaw: Energy Market Agency; 2009. [54] Lopez de Haro S, Sanchez Martin P, de la Hoz Ardiz JE, Fernandez Caro J. Estimating conjectural variations for electricity market models. European Journal of Operational Research 2007;181. [55] Díaz CA, Villar J, Campos FA, Reneses J. Electricity market equilibrium based on conjectural variations. Electric Power Systems Research December 2010; 80(12):1572e9. [56] Lorenz. Gospodarka we˛ glem kamiennym energetycznym. Kraków: IGSMiE PAN; 2010.
ski Z. Konkurencyjno
s
[57] Grudzin c wytwarzania energii elektrycznej z we˛ gla brunatnego i kamiennego. Polityka Energetyczna. Kraków: IGSMiE PAN; 2010. t. 13, z. 2. [58] Brook A, Kendrick D, Meeraus A. GAMS users’ guide, release 2.54. San Francisco: The Scientific Press; 1992. [59] Dirkse SP, Ferris MC. The path solver: a non-monotone stabilization scheme for mixed complementarity problems. Optimization Methods and Software 1995;5:123e56. [60] Ferrris MC, Munson TS. Complementarity problems in GAMS and the PATH solver. Journal of Economic Dynamics and Control February 2000;24(2): 165e88.

6644


ski / Energy 36 (2011) 6634e6644 J. Kamin

[61] Al-Agtash SY. Supply curve bidding of electricity in constrained power networks. Energy July 2010;35(7):2886e92. [62] Carraretto C, Zigante A. Interaction among competitive producers in the electricity market: An iterative market model for the strategic management of thermal power plants. Energy December 2006;31(15):3145e58.

[63] Most D, Perlwitz H. Prospects of gas supply until 2020 in Europe and its relevance for the power sector in the context of emission trading. Energy 2009;34:1510e22. [64] Pa1ka P. Uwolnienie cen detalicznych na obecnym rynku energii elektrycznej. Rynek Energii 2011;2(93).