Vous êtes sur la page 1sur 27

The European Electricity Market Does Liberalisation Bring

Cheaper and Greener Electricity?

Claudia KEMFERTa, Wietze LISEb, and Robert STLING c

1. DRAFT October 2003

Scientific Pool of Environmental Economic Disciplines (SPEED), Oldenburg University, Oldenburg, Germany.

Institute for Environmental Studies, Faculty of Earth and Life Sciences, Vrije Universiteit, Amsterdam, The Netherlands

Stockholm School of Economics, Stockholm, Sweden

Address for correspondence:


Claudia Kemfert
University of Oldenburg
D- 26111 Oldenburg
E-mail: kemfert@uni-oldenburg.de
Tel: +49 441 798 4106
Fax: +49 441 798 4101

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Abstract
This paper investigates the impacts of the liberalisation of the European electricity market. We apply a game theoretic modelling framework in order to assess the economic
consequences of different market behaviour. It turns out that, because of different developments of European countries, some countries benefit substantially by strategic market
behaviour whereas others could also profit from more competition. The effects on the
environment are ambiguous.
Keywords:

Electricity market liberalisation, Game theoretic model, Environmental


effectiveness, Europe

JELclassification: C7, D2, Q4, R3

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

1. Introduction
The liberalisation of electricity marketsthe introduction of competition, the reduction
of external, particularly political interferences and adjustments, and the opening of the
market for new providersis a worldwide phenomenon. Although the reasons for an
opening of markets vary from country to country, the main goal apart from higher production efficiency is to offer the customers lower electricity prices.
Though only a few countries in the world have accomplished a complete liberalisation, it
can be observed, however, that most countries aim for a complete opening of the electricity market in the near future (see Figure 1). In Europe, all EU member states shall
liberalise the electricity market according to the 1997 directive of the European Commission (Directive 96/92/EC). The directive provides that European electricity markets
should already have been opened up to an average of 25 percent in 1999.
The directive is, however, converted differently into actual policy in different countries,
and the progress of liberalisation of electricity markets in Europe vary between countries. In Germany the market was liberalised in 1999, whereby the markets in Norway,
Sweden and United Kingdom had already been completely opened up by that time. Austria and Denmark have liberalised their electricity markets almost completely as well
(see Table 1). Spain is likewise aiming at an imminent opening of the market. France and
Italy have not yet decided when they intend to open the market for external competition.
These countries are characterized by the predominant position of a few electricity produces, i.e. a rather uncompetitive monopoly and/or duopoly prevail (France: EdF, Italy:
Elettrogen and Enel).
This unequal distribution of market opening and liberalisation of the electricity markets
in Europe involve some competition distortions some utilities already face complete
competition, whereas others can continue operating in a monopolistic position. Since
utilities have to compete with each other after opening of the market, providers, in order
to survive, need to alter their behaviour. In Germany, for example, utilities reacted very
dynamically after the liberalisation of the electricity market in 1999 by firm mergers and
strategic behaviour. A rise of the market shares of certain producers might lead to a
rather uncompetitive market structure, which will not reduce, but rather increase electricity tariffs.
Whether an electricity supplier is able to convert strategies in the electricity sector depends on the market situation, in particular on the dominant market conditions. Thus the
market entrance conditions on the different levels of the current market (production,
trade (and selling)) play a crucial role.
Furthermore, also electricity trading options can offer additional incentives for the practice of market power, unless uniform price structuring for tradable electricity are created.
In Germany for example, a federation agreement regulates the prices for the power trade.
However, it has been observed in the past that due to strategic market behaviour, providers for third party access of extraneous electricity were completely delayed or refused. In
its second benchmark report the European Commission criticizes that competition distortions and market power can arise through high entrance net access fees, which obstruct
3

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

the entrance of new providers. The different market opening degrees diminishes the advantages for the customer.
Whether the liberalisation of the European electricity markets involves a sustainable, i.e.
environmental friendly development, remains to be observed. The current market liberalisation, and the political decisions in some European countries on nuclear energy and
renewable energy law, can lead to increased current production costs, which can involve
an increased import of comparatively low-cost electricity. Cheap electricity is not necessarily beneficial from an environmental point of view as it is often produced by depreciated production plants like nuclear or coal An environmental beneficial situation would
increase the share of renewable energy like wind, hydro or biomass technology. Electricity from nuclear energy is to be judged positively from a climatic point of view, since it
sets free only few greenhouse gases like the main greenhouse gas carbon dioxide. However, nuclear energy poses other dangers and risks to the environment, which usually are
not included in the production calculation. As in the current liberalised market, green
energy involves higher production costs, and is less competitive. Hence, a harmonisation
of environmental laws seems to be necessary in Europe.
Newberry (2002, 2002a, 2002b) studied potentials and opportunities for European utilities. Day and Bunn (1999) investigated these aspects by a game theoretic model of market power and strategic actions of firms in the UK. Bower and Bunn (1999) assess trade
opportunities within a pool versus a bilateral trade system in the electricity market of the
UK. Admundsen and Bergman (2002) studied these issues for the Norwegian and Swedish power market. Here, transmission and transport pricing plays a crucial role.
Experiences in Scandinavia and the UK suggest that a uniform tariff is preferred over
distance charges. Moreover, market opportunities and grid owners significantly influence
trade. Dawson and Shuttleworth (1997) studied transmission pricing in Norway and
Sweden. Green (1997) examined this effect for the UK. Cardell et al (1994) investigated
the negative effects of market power and transmission constraints on trading by an imperfect competition model for North American electricity suppliers.
Different non-cooperative games within various markets have been examined by diverse
authors. Murphy et al (1986) demonstrate how mathematical programming approaches
can be used to determine oligopolitistic market equilibria. Salant and Shaffer (1999) illustrate the theoretical impacts on production and social welfare via two-stage CournotNash equilibrium solutions, where learning by doing and investments in R&D determine
marginal costs of identical agents differently. For Europe, Jing-Yuan and Smeers (1999)
have modelled an oligopolistic electricity market with a sophisticated game theoretic
model, calculating the Nash equilibria. More generally, Helman et al (1999) investigated
different kind of trade options and strategic price setting within the electricity market.
Stern (1998) investigates the liberalisation of the European gas market.
Bower et al (2001) have simulated the liberalised German electricity market by an agentbased model. They conclude that mergers increase market power, increasing the electricity prices. Their model is very sensitive to the out-phasing of expensive oil-fired plants,
nuclear energy, or to closing the borders to imports of (cheap) electricity. In all these instances, prices jump up considerably. Bigano and Proost (2002) conducted a four country study (France, Germany, Belgium, the Netherlands) which is linked through electric4

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

ity trade. The environmental impacts are quantified, where a 3-stage game is calculated
in a partial equilibrium framework. They compare strategic action with perfect competition to conclude that phasing out nuclear energy leads to a substantial decrease in social
welfare.
In a liberalised electricity market, electricity suppliers can act strategically, which influence electricity prices, due to changing market shares in favour of large firms. Furthermore, mergers can become attractive, as it increases the electricity price and hence profits. While enhancing competition in the electricity market, strategic behaviour determines the structure of the market and energy supply network (see also Kemfert, 1999,
and Kemfert and Tol, 2000).
In this paper, strategic behaviour of energy suppliers and their impacts on the economic
and environmental situation in the liberalised European electricity market is studied with
the game theoretic modelling tool EMELIE (Electricity MarkEt Liberalisation In
Europe).1 EMELIE is calibrated to the main European energy suppliers, which are linked
by capital flows. These firms produce electricity through different technologies, considering others and their own capacities, and variable production costs. Within this context,
the analysis focuses on the impacts of strategic action by firms, which is compared to the
case without strategic action, i.e. when the market is fully competitive. We investigate
the electricity prices and the environmental effectiveness of applied technologies.
This paper is outlined as follows. Section 2 highlights some important aspects of
Europes electricity market structure. Section 3 gives a general introduction to the computational game theoretic model EMELIE, while the Annex shows a more detailed
mathematical formalisation. The required data and the calibration of the model to the
European liberalised electricity market are presented in Section 4. Section 5 discusses
the main model results for seven European countries, namely Denmark, Finland, France,
Germany, the Netherlands, Norway and Sweden. The final section concludes.

2. Europes Electricity Market Structure


As the benchmark report of the European Commission of March 2003 testifies, the
European electricity market can be characterised by increased market opening, improvements of unbundling of net owners and more transparent regulation methods.2
Whereas Italy and United Kingdom registered decreased electricity prices for large consumers, also Austria, Germany and the Netherlands observed increasing activities of customers. However, as not all countries have liberalised their electricity markets completely, some unsolved difficulties still emerge, as the degree of unbundling, increased
market shares of dominant utilities in some European countries and the lack of infra-

A first version of EMELIE is applied to study economic impacts of the German electricity market in Lise et al (2003). First application to the European market is given by Kemfert et al
(2002).
2
See European Commission (2003).

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

structure of inter country electricity trade. Considerable electricity price decreases could
also be observed in Austria, as Figure 3 demonstrates.
Since the market conditions for the current providers are still very different in the individual countries in Europe, there is so far no clear tendency to observe in the development of the electricity market. In Germany the liberalisation of the electricity market
first led to sinking electricity tariffs, particularly for the main customers, due to increased
competition. For the private customer range the prices reduced first. However so far,
only a few private customers in Germany have changed provider (see Table 1). Therefore, within the private customer range in Germany, the electricity tariffs rose once
again. In England, however, private customers frequently changed provider due to strong
electricity tariff variations.
Through various developments of the electricity market in individual European countries
and the additional competitive pressure on the individual providers, strong strategic behaviour of individual providers increases. Mergers of large power suppliers in Germany
reduced and rather obstructed competition, since the free decentralised market access
also is decreased for small electricity providers. Although France so far has opened its
electricity market, the largest French provider EdF dominates the electricity supply in
France and increasingly also in Europe. Because the nationally controlled giant pursued
a strong policy of expansion overseas, it is however hardly possible for current European
providers to expand into the French market. Therefore the European commission has already demanded extra time to regulate the European market as uniformly as possible and
to decrease market power by an independent adjustment authority.

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Table 1: Liberalisation of the Electricity Market in Europe


Country

Austria

Percent
of liberalisation
100%

Date of complete liberalisation


2003

Belgium
Denmark
Finland
France

35%
90%
100%
30%

Germany

100%

2007
2003
1997
Discussion
not ended
1999

Main providers

EVN, Verbund, Wiener


Stadtwerke
Electrabel
SK Power Company
Fortrum, Ivo Group
EDF

Market share
of the main
providers
68%

Consumers
who changed
providers
5-10%

97%
75%
54%
98%

5-10%
N/A
30%
5-10%

Bewag, E.On, EnBW,


63%
10-20%
RWE, Veag
Greece
30%
Not discussed AEH (public company)
100%
None
Ireland
97%
2007
ESB
97%
30%
Italy
35%
Not discussed Elettrogen, Enel
79%
Less than 5%
Luxemburg 50%
2007
Cegetel
90%
N/A
Netherlands 33%
2003
Essent, Nea
64%
10-20%
Portugal
30%
Not discussed EDP
85%
Less than 5%
Spain
45%
2003
Endesa, Hidroelectrica
79%
Less than 5%
del Cantabrico, Iberdrola, Union Fenosa
Sweden
100%
1998
Sydkraft, Vattenfall
77%
N/A
UK
100%
1998
British Energy, Innogy,
44%
80%
Powergen, Scottish and
Southern Energy, Scottish Power
Source: Benchmark Report of the European Commission, see European Commissions (2003)

In Germany, the liberalisation of the electricity market first led to substantial electricity
price reductions, but increased again because of firm mergers and increased market
shares. Lise, Kemfert and Tol (2002) report on this and provide detailed model simulations.
The European Commission criticised especially the net access price developments in the
German market. Whereas the differences between net purchase and sales prices are at a
constant low level (with few exceptions in peak times) in Austria and United Kingdom,
the German net access sales prices are permanently at a very high level. In addition, the
purchase prices are very low in Germany so that the difference between purchase and
sales price is continually at a very high level, independently of hourly differences, see
Figure 4.

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

3. The Game Theoretic Modelling Approach


We investigate the market developments in seven European countries using the game
theoretic model EMELIE (Electricity MarkEt Liberalisation In Europe). EMELIE can be
characterised as a small numerical model, with which market behaviour by firms in oligopolistic markets can be investigated. Energy suppliers in EMELIE produce electricity
through different technologies. One player can own several power plants, of which total
capacity is considered, as well as variable production costs. Investment decisions are not
considered here; we limit ourselves to the study of the economic and environmental impact of liberalisation with the present capacity structure.
The model is calibrated by providing the electricity retail price and the actual demand in
a particular base year. In this paper we take the year 2000 as the base year. In the calibration of the model, production costs are minimised to see whether it is possible to meet
the required demand. Below we refer to the results from this calibration as the reference
case (REF).
After calibration the EMELIE model use three different sets of assumptions about the
behaviour of the producers in the model to compute the market outcome in equilibrium.
Besides data on electricity producers, price elasticities of demand and transmission grid
capacity are specified exogenously. EMELIE then determine national electricity prices,
marginal electricity production costs, and produced and traded electricity per technology
and firm.
The first case uses the assumptions of the perfectly competitive market, i.e. that firms are
price takers and act as if they couldnt influence the market price. This case is referred to
as the perfect competition case, COMP. The perfectly competitive outcome is equivalent
to Bertrand oligopoly competition, i.e. when firms set prices taking into account the actions of the other firms.
In contrast to the perfectly competitive situation, in the second case producers are assumed to act strategically. Each player decide on production quantity taking into account
the strategic choices of the other players. This is the Cournot-Nash oligopoly model,
which is characterised by mutual, strategic reactions by market players. This result leads
to a Nash equilibrium where the strategies of all market actors are best responses to the
actions of all other market players. However, a so-called competitive fringe consisting of
the total sum of small decentralised production units has been included, and this fringe is
assumed to always behave as price takers. Strategically behaving firms can influence
prices by changing production. We refer to this as the STRA case.
In the Stackelberg game, one firm, in this case the one with the largest production capacity, is a leader and moves first, while other firms are followers and take the action of the
leader as given. The followers behave just like the STRA case, except for the competitive fringe who always behaves as in the COMP case. The leader chooses its output
knowing ultimately the outcome will lie on the followers reaction curve, which means
that the followers reaction curve is a constraint for the leader. The leader chooses its
output to maximise profit given this constraint. We refer to this case as the STACK case.
8

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

The Annex provides a more detailed description of the EMELIE model.


The model is implemented in GAMS and formulated as a mixed complementarity problem (MCP). This is both a fast and efficient way to solve the model for different scenarios and with different data sets.

4. Calibration of EMELIE to the European Electricity Market


As was explained in the previous section, the EMELIE model is a game-theoretic computable model of the wholesale electricity market. The model uses different assumptions
about market structuresBertrand, Cournot-Nash and Stackelberg oligopolyto compute the market outcome in terms of prices, quantities and technology use. The current
version of the EMELIE model is static and considers international trade exogenously,
namely by modelling each neighbouring country as a separate player with a production
capacity equal to the transmission capacity from that country. The model is run separately for each country.
The supply side of the model consists of a number of electricity producers that produce
and sell electricity. For each producer the production capacity of different production
technologies are specified. To account for the great number of small decentralised production units, a competitive fringe is added to the model. As their sizes are small, they
always act as price takers in the model.
The demand side of the model consists of one sector demanding the electricity produced,
as we generally observe only a single power exchange market within a country. This
market trades electricity measured in terawatt hours (TWh). As we are mainly interested
in the economic consequences of liberalisation, we consider trade over the period of one
year. It is not possible to associate emissions to a model that only considers the possibility to meet peak demand. Hence, the model outcome shows to which extent demand can
be met within the current capacity structure.
The model allows for many different production technologies. At this stage 12 different
production technologies have been considered. Conventional thermal power technologies
are nuclear (N), coal (C), gas (G), lignite (L) and oil (O). There are five different types of
combined heat and power production (CHP): gas (CHP-G), coal (CHP-C), oil (CHP-O),
biomass (CHP-B) and other fuels (CHP-X). Finally, hydro (H) and wind power (W) are
also available technologies in the model.
We calibrate the EMELIE model by considering data for the benchmark year 2000: production capacities of the largest producers, variable production costs for different technologies, transmission capacities between countries, together with wholesale market
price and demand data.
The data used in this report is from the base year 2000 and includes the following countries: Denmark, Finland, France, Germany, the Netherlands, Norway and Sweden. A
brief summary of the total production capacities in these countries is given in Table 2.

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Table 2 Electricity production capacities in 2000


(TWh /
year)

Nuclear
Coal

Denmark

Finland

France
N

Germany

9.2 21.3 16.7 403.6 23.8 167.9 35.6

16.8 2.6 16.0

88.8

13.1 178.9 26.8


2.4

99.3

Netherlands

Norway

Sweden

3.8

5.4

9.1

68.1

3.5

34.6

3.9

17.3

0.4

Lignite

1.5

Gas

2.7

6.3

0.4

13.2

18.5 157.7 10.2

61.2

0.5

1.5

Oil

2.8

8.7

1.8

85.6

10.3

64.1

3.5

8.5

2.5

22.8

1.4

CHP-G

8.6

0.2

8.1 0.06

0.2

10.2

3.5

39.8

2.1

0.6

5.8

CHP-C

7.1

1.0

6.6

0.2

1.3

52.1

2.9

2.0

2.5

5.0

CHP-O

0.3

0.7

0.3

0.06

2.6

0.07

0.4

2.9

0.1

CHP-B

0.2

4.7

0.2

0.05

5.4

0.3

2.1

0.6

CHP-X

1.0

6.5

0.4

29.9

0.8

32.9

1.2

0.9

0.6

4.5

1.1

Hydro

0.03 15.7 12.5 6.8

72.6

38.9

37.4

33.9

0.3

Wind

4.4 0.06 0.08 0.05

0.2

0.02

0.7

1.8

1.9

Total

37.0 37.3 91.4 27.0 693.9 109.5 803.6 119.6 155.7 27.5 143.2 30.7 168.5 62.2

22.1 142.3 8.2


-

0.03

1.1

64.4 23.1
0.5

2.4

Notes: N shows total domestic capacity and I refers to import capacity.

There are differences in available technologies between countries, and not all technologies are present in each country. For each country and technology variable production
costs have been specified, which are shown in Table 3.

Table 3 Electricity production costs in 2000


(/MWh) Denmark Finland France Germany Netherlands Norway
Sweden
Nuclear
14.22
2.96
5.00
11.19
5.00
SWE
4.32
Coal
14.88
DEN
15.11
FRA
23.00
DEN
DEN
Lignite
FRA
GER
15.65
GER
Gas
18.29
30.52
44.45
34.28
41.73
SWECHP Oil
18.61
46.91
48.46
36.70
53.83
SWE
79.69/54.40
CHP-G
10.54
14.99
GER
20.21
21.90
DEN
29.68
CHP-C
8.61
9.10
GER
9.64
DEN
21.94
CHP-O
9.30
23.33
GER
20.32
SWE
35.22
CHP-B
11.76
SWE
SWE
SWE
19.06
SWE
19.15
CHP-X
SWE
16.64
20.98
10.00
SWE
26.50
Hydro
1.28
0
12.93
1.30
9.08
0
3.55
Wind
1.28
0
0
1.30
9.08
DEN
0
Notes: When a production cost estimate for another country has been used, this is indicated by
the first three letter of the country from which the cost estimate is borrowed. SWECHP indicates that the Swedish cost estimate for CHP has been used. The oil cost estimate for Sweden
refers to gas turbine and condensing, respectively.

10

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

There are differences in variable production costs between countries, which is the result
of different fuel and production taxes in different countries. These differences also reflect that it is very difficult to get exact production cost estimates, although it should be
noted that relative costs between different technologies are roughly the same in all countries. In some cases, national cost estimates were not available, and then cost estimates
for some other country has been used, as indicated in Table 3.
Capacity and production cost data and the price elasticity of demand are crucial model
data and parameter. Electricity consumers are not very price sensitive. Therefore, we assume a rather moderate price elasticity of demand of 0.400 for five of the countries,
while we use slightly smaller price elasticities for Denmark and Netherlands (0.295 and
0.290, respectively), based on average values found in the literature (see for example
Andersson, 1997, Brubakk et al, 1995, Grohnheit and Larsen, 2001 and Koopmans et al,
1999).

5. Country Results and Sensitivity Analysis


The EMELIE model has been run for seven countries and four cases. Table 4
summarises the results by stating the number of (strategically acting) firms, the number
of trading partners, and the equilibrium prices and quantities for the four different cases:
Table 4 Model results for seven countries
n

POP
(mil)

REF
p

COMP
Q

STACK
p
q

STRA
p

Denmark
2 3
5 12.75
32.83 10.54
34.73 14.20
31.81 14.29
31.75
Finland
2 3
5 14.88
76.41 14.88 76.41 16.64 73.07 16.64 73.07
France
1 6
60 20.81 410.67 12.93 496.78 24.85 382.55 48.46 292.85
Germany
4 8
83 15.19 477.00 15.11 478.01 20.33 424.53 20.36 424.28
Netherlands 6 2
16 39.65 100.60 29.11 110.03 38.71 101.30 38.71 101.30
Norway
6 4
5 12.25 110.92 5.23 155.87
5.50 152.85 5.50 152.85
Sweden
6 5
9 14.26 135.46 9.83 157.20 14.98 132.82 15.78 130.08
Notes: n refers to the number of modelled firms in each country, c refers to the number of modelled neighbouring countries, p refers to the price in euros per MWh, and q refers to national
supply in TWh.

A number of insights can be derived already from Table 4 . For instance the following
relation holds:

p COMP p STACK p STRA and p COMP p REF


On the one hand, we find that the reference price for Finland and Germany are (nearly)
equal to the price in perfect competition. This means that the actual production capacity
and costs are so stringent that they are just enough to meet demand in the fully competitive market. It also indicates that in these two countries almost perfect competition or full
liberalisation is already reached. Especially in Germany, we observed huge changes due
to increased competition within the last two years.
11

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

On the other hand, we find that the reference price is even higher than Cournot price for
the Netherlands, Norway and Sweden. Therefore, further liberalisation of the markets in
these countries should lead to lower prices. It also indicates that in Norway, in the Netherlands and in Sweden firms choose the market price far above the competitive price
(marginal cost) in order to increase profits. However, in comparison to the full competition case no additional profits could be reached by becoming a Stackelberg leader in
these three countries.
In Norway firms obtain prices far above the full competitive price, as the COMP,
STACK and STRA price is less than half the reference price (see Table 4). The main explanation for this is that the price difference among competing technologies is quite low.
Furthermore, Norway has a high hydro production capacity. First, hydro power capacity
varies from year to year, and the measure that has been used in the model is the actual
production in 2000, when hydro production was very high. Second, transmission capacity between Sweden and Norway is very high, and Norway can therefore benefit from
importing cheap hydro and nuclear power from Sweden. The same reasoning holds for
Sweden.
In Finland we can see that firms already have chosen the perfectly competitive price.
Full liberalisation of the electricity market in Norway has already brought a fully competitive situation. However, firms could increase their gains by acting strategically.
Finally, the model finds very high prices for the Cournot equilibrium in France as compared to other countries. The Cournot price is about four times the competitive price,
which is of course a result of the dominant, near monopoly, position of EdF. These high
prices are caused by a (strategic) drop in production by EdF. Possibly, the consumer
price elasticity of French consumers is underestimated here, as when this value is raised
to 0.8, we no longer find this extreme result (pSTRA=22.33 /MWh). It could also be
noted that despite the predominant position of EdF, the electricity price in France is not
particularly high (except for the Cournot price). This is most easily explained by the
large nuclear power capacity, which has a low marginal cost of production.
The price of electricity in the Netherlands is very high compared to other countries in
2000, which is a result of high oil and gas prices in year 2000. The results for the Netherlands are similar to the results for other countries. The Cournot price of 38.71 euros
might appear high, but it is about 33 percent higher than the competitive price, which is
similar to most other countries.
In Finland, the Netherlands and Norway, the STACK outcome is equal to STRA, while
it is interesting to know that in the other countries the Stackleberg leader always increases profits at the cost of lowering the profits of all other firms. According to economic theory, the Stackelberg equilibrium price is expected to be in between the Cournot
and perfect competition price. The model run results are therefore in line with economic
theory. Economic theory also predicts that the Stackelberg leader is generally better off
in the Stackelberg equilibrium as compared to the Cournot equilibrium. This is also what
the model results suggest for the various countries studied (the result on the profits is not
shown in this paper).
Besides, we can also study the differences in competitiveness within the various countries by calculating the HHI (the formula can be found in the appendix).
12

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Table 5 The Hirschmann-Herfindahl index for the seven countries and four model runs.
Denmark
Finland
France
Germany Netherlands Norway
Sweden
REF
1520
2063
9048
1667
1066
1575
1888
COMP
1405
2096
7414
1651
985
894
2115
STACK
1315
1237
3871
926
885
892
1565
STRA
1314
1237
1432
923
885
892
1495
Note: The Fringe which consists of many small firms is excluded from the HHI calculation.

Table 5 clearly shows that France is close to a monopoly. This may also explain the extreme price increase once the dominating firm starts to act strategically. Once a firm exploits its market power, the HHI value naturally drops due to the lower supply. Furthermore, it is interesting to see that besides France, Finland and Sweden have particularly
high market concentrations, as their HHI values are above the threshold of 1800 in the
reference situation.
The comparison of the results for different countries showed that in all studied countries
(Cournot) strategic behaviour of firms leads to higher electricity prices and lower production quantities than in both the reference and perfect competition cases. Because of
the monopolistic electricity market situation in France, strategic behaviour leads to very
high prices and a substantially reduced electricity production. In Norway, we find very
little difference between different scenarios. This is primarily because utilities always
produce at full capacity in all scenarios, as electricity can be produced at zero cost (hydro and wind). Therefore, the Norwegian electricity market is resistant against any market changes towards more competition or strategic actions.
These broad statements about the model results give a general picture of the performance
of the EMELIE model with data for year 2000. A more detailed picture of the results is
given in the next section, where the results are analysed individually for each country.

5.1 Denmark
In contrast to other Nordic countries, Denmark liberalised its electricity market slowly
but reaches presently a nearly full market opening. There are two large producers in
Denmark, Elsam and Energi E2, which control 40 and 34 percent of total national generation capacity, respectively.
There are two peculiarities with the Danish electricity market that are noteworthy. First,
the market has a natural regional division. The electricity grid in Denmark is divided between the eastern and western parts of the country, and in many respects each part is
more integrated with the surrounding countries (Sweden, Norway and Germany) than
with each other. This situation, however, does not influence the model results since
Denmark is treated as one integrated market in the model. Secondly, CHP is the second
most important electricity generation technology after conventional coal combustion.
The higher efficiency of CHP is, like for other countries, captured by lower technology
prices in the model. Table 6 presents the results for Denmark.

13

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Table 6 Model results for Denmark


p

Market shares, %
Technology shares, % Marginal technologies
(Elsam / E2 Energi /
(CHP-G / CHP-C /
Fringe / Sweden /
CHP-O / H / W)
Norway / Germany)
CHP-G
REF
12.75 20 / 6 / 24 / 21 / 26 / 4 15 / 24 / 1 / 46 / 13
CHP-G
COMP
10.54 19 / 8 / 26 / 20 / 24 / 4 20 / 23 / 1 / 44 / 13
CHP-G, CHP-C, CHP-O
STACK 14.20 12 / 8 / 30 / 20 / 27 / 5 20 / 19 / 0.1 / 48 / 14
CHP-G, CHP-C, CHP-O
STRA
14.29 12 / 8 / 30 / 20 / 27 / 5 20 / 19 / 0.1 / 48 / 14
Notes: The market share of a player is defined as the production of the player divided by the total
production of all players. Technology shares denote the share of a certain technology as a
percentage of total production. Marginal technologies are technologies for which only a part
of total capacity is used in production. The Stackelberg leader is in bold font.

Table 6 shows that hydro power production in Sweden and Norway account for almost
half of the supply in Denmark. It should be noted, however, that international trade is
exogenous in the model, which means that Danish imports do not reduce the supply in
Sweden and Norway. Table 6 also verifies the important role of co-generation (CHP) in
Denmark CHP accounts for almost half of total supply and it is the marginal technology in all cases. This means that the outcome hinges critically on assumptions about
CHP capacities and production costs.

5.2 Finland
The Finnish electricity market is fully liberalised since 1997. The market is dominated
by the state-owned producer Fortum which controls about 35 percent of total national
electricity generation capacity. The second largest player, PVO, is owned by electricityintensive industry and controls around 20 percent of total capacity. Decentralised production, the competitive fringe, is quite substantial and accounts for about 40 percent of
total capacity. The electricity production technology is mixed, although nuclear is the
most important technology.
Table 7 Model results for Finland
p

REF
COMP
STACK

14.88
14.88
16.64

Market shares, %
(Fortum / PVO / Fringe /
Sweden / Russia / Norway)
36 / 18 / 16 / 17 / 13 / 1
36 / 18 / 15 / 17 / 13 / 1
24 / 14 / 31 / 17 / 14 / 1

STRA

16.64

24 / 14 / 31 / 17 / 14 / 1

Technology shares, %
(N / C / CHP-G / CHP-C /
CHP-X / H / W)
48 / 18 / 0 / 9 / 0 / 25 / 0.2
48 / 18 / 0 / 9 / 0 / 25 / 0.2
51 / 3 / 8 / 8 / 5 / 26 / 0.2
51 / 3 / 8 / 8 / 5 / 26 / 0.2

Marginal technologies
C
C
C, CHP-G,
CHP-C, CHP-X
C, CHP-G,
CHP-C, CHP-X

Notes: See notes under Table 6.

5.3 France
France has liberalised its electricity market gradually. Although the market is liberalized
in some respects, the state-owned producer EdF dominates the market with some 85 per14

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

cent of total capacity; a near monopoly situation. The dominating technology is nuclear
power, accounting for a major part of total electricity production in France. Additionally,
France is a net-exporter of electricity with net exports of about 70 TWh in 2000.
The result in the STRA case is quite different from the others, as can be seen from Table
8. In the Cournot equilibrium EdF restricts output so that its market share drops from 86
to 36 percent! Interestingly, in this scenario the market share of the competitive fringe is
much higher than in the other scenarios producing a higher share of electricity production with a completely different mix of technologies, see Table 8.
Finally, we can also observe that the Stackelberg price is twice the competitive price,
which is a result of the Stackelberg leader, EdF, being so large compared to the other
players. EdF is able to increase profits by acting as a leader. Then the consumer is much
better off than under STRA as the price is halved in STACK.
Table 8 Model results for France
p

Market shares, %
(EdF / Fringe / Switzerland / Italy / Belgium /
Germany / UK / Spain)
REF
20.81 95 / 0 / 0 / 0 / 3 / 1 / 1 / 1
COMP 12.93 86 / 2 / 5 / 2 / 2 / 1 / 1 / 1
STACK 24.85 62 / 18 / 7 / 2 / 3 / 4 / 3 / 2
STRA
48.46 36 / 37 / 9 / 3 / 4 / 5 / 4 / 3
Notes: See notes under Table 6.

Technology shares, %
(N / C / L / G / O / CHP-G /
CHP-C / CHP-O / H / W)

Marginal
technologies

100/ 0/0/0/ 0/ 0/ 0/ 0/ 0/0.04


84/ 0/0/0/ 0/ 0/0.2/16/ 0/0.04
68/12/1/0/ 0/0.1/0.3/ 8/12/0.04
44/15/1/4/10/0.1/0.4/10/15/0.05

N
H
N, C, H
N, C, G, O, H

5.4 Germany
Germany is one of the European countries that have fully liberalised their electricity
markets. There are four large electricity producers in Germany, namely Vattenfall, EON, EnBW, RWE and a competitive fringe having 20% of total national capacity. Vattenfall is the largest electricity producer, owning 32% of total national capacity.
Table 9 Model results for Germany
p

Market shares, %
(E-ON / EnBW / RWE / Vattenfall /
Fringe / Austria / Benelux / Switzerland /
France/Denmark/Poland/Sweden/Central)
REF
15.19 17 / 19 / 6 / 31 / 12 / 3 / 2 / 3 / 4 / 1 /0/1/2
COMP
15.11 17 / 19 / 6 / 30 / 12 / 3 / 2 / 3 / 4 / 1 /0/1/2
STACK 20.33 18 / 10 / 9 / 18 / 21 / 4 / 2 / 3 / 5 / 3 /2/1/5
STRA
20.36 18 / 10 / 9 / 18 / 21 / 4 / 2 / 3 / 5 / 3 /2/1/5
Notes: See notes under Table 6.

Technology shares,
%
(N / C / L / CHP-C /
CHP-X / H / W)
41 /26 /0 /11 /7 /15 /1
41 /26 /0 /11 /7 /15 /1
46 /15 /2 /13 /8 /16 /1
46 /15 /2 /13 /8 /16 /1

Marginal
technologies
C
C
N, C, L
N, C, L

As we have already seen from Table 3, in Germany firms are already producing electricity in an almost fully liberalised and competitive market situation. Both market
shares of producing firms and technology shares are almost identical in the reference
15

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

and competition situation. Strategic behaviour leads to different shares of firms and
technology mix.
5.5 The Netherlands
In the Netherlands, gas, exploited from Dutch sources, is the most important fuel used in
electricity production. In 2003, the Netherlands has opened up its electricity market to 33
percent. Six different electricity producers have been modelled for the Netherlands together with a competitive fringe consisting of around 20 percent of total capacity. The
largest firm is Electrabel which controls around 25 percent of total capacity. Essent and
Reliant are the second largest firms, each controlling about 20 percent of national capacity.
Table 10 Model results for the Netherlands
p

Market shares, %
(Electrabel / Reliant / E-ON / Essent
/ Delta / NUON / Fringe / Belgium /
Germany)
REF
39.65 4 / 9 / 7 / 21 / 4 / 1 / 28 / 5 / 21
COMP 29.11 5 / 10 / 8 / 19 / 5 / 3 / 26 / 5 / 20
STACK 38.71 6 / 11 / 9 / 12 / 6 / 3 / 28 / 5 / 21
STRA
38.71 6 / 11 / 9 / 12 / 6 / 3 / 28 / 5 / 21
Notes: See notes under Table 6.

Technology shares, %
(N / C / CHP-G /
CHP-B / H / W)

Marginal
technologies

9 / 24 / 38 / 5 / 22 / 2
8 / 31 / 35 / 5 / 20 / 2
9 / 25 / 38 / 5 / 21 / 2
9 / 25 / 38 / 5 / 21 / 2

C
C

5.6 Norway
Norway liberalised its electricity market earlier than all other countries studied here, and
it is also the home country of the Nordic electricity exchange Nordpool. Norway is also
special in that almost all of the generation capacity consists of hydro power. This means
that generation capacity varies from year to year depending on hydrological conditions
in 2000 Norway produced a huge amount of hydro power and it is perhaps not a very
representative year. The dominance of hydro power therefore makes it difficult to choose
a correct electricity generation capacity figure since it varies from year to year. The variable production cost of hydro power is set to zero in the model.
The dominating electricity producer in Norway is the state-owned company Statkraft
which controls about 30 percent of total capacity. The other five players that have been
modelled are rather small, controlling around 58 percent of total capacity. The competitive fringe is therefore quite large, accounting for almost 40 percent of total capacity.

16

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Table 11 Model results for Norway


p

Market shares, %
(Statkraft / Oslo / Norsk Hydro / Agder / BKK
/ Lyse / Fringe / Sweden / Denmark / Finland /
Russia)
REF
12.25 35 / 8 / 10 / 8 / 7 / 6 / 21 / 5 / 0 / 0.1 / 0.0
COMP
5.23 25 / 6 / 7 / 6 / 5 / 4 / 36 / 10 / 1 / 0.2 / 0.3
STACK
5.50 26 / 6 / 7 / 6 / 5 / 4 / 36 / 9 / 1 / 0.2 / 0.3
STRA
5.50 26 / 6 / 7 / 6 / 5 / 4 / 36 / 9 / 1 / 0.2 / 0.3
Notes: See notes under Table 6.

Technology
shares, %
(N / H / W)

Marginal
technologies

0 / 100 / 0
6 / 94 / 1
4 / 96 / 1
4 / 96 / 1

N
N

5.7 Sweden
Sweden has like the other Nordic countries reached far in the process of liberalisation.
The dominating player is state-owned Vattenfall which controls more than 40 percent of
total capacity. The second largest player is Sydkraft, which controls 20 percent of total
capacity. The third largest player in 2000 was Birka Energi with 13 percent of total capacity. Later the Finnish producer Fortum has aquired Birka Energi and shares in other
companies so that Fortum is now the third largest producer in Sweden. In total, six players have been modelled together with a fringe consisting of 17 percent of total capacity.
Nuclear and hydro power are the most important production technologies in Sweden. In
2000, hydro power production was very high and actual production exceeded the average
production, which has been used as a capacity measure here.
In Sweden the Stackelberg equilibrium reduces the price under strategic behaviour with
5%. The situation resembles that of France with a large, predominant producer, Vattenfall, which results in a quite substantial first-mover advantage.
Table 12 Model results for Sweden
p

Market shares, %
(Vattenfall / Sydkraft / Birka / Fortum /
Skellefte / Graninge / Fringe / Norway /
Finland / Denmark / Germany / Poland)
REF
14.26 51 / 8 / 8 / 4 / 2 / 2 / 4 / 15 / 3 / 2 / 1 / 0
COMP
9.83 44 / 16 / 11 / 4 / 2 / 1 / 4 / 13 / 3 / 2 / 1 / 0
STACK 14.98 34 / 19 / 13 / 4 / 2 / 2 / 5 / 16 / 3 / 2 / 1 / 0.3
STRA
15.78 29 / 20 / 14 / 4 / 2 / 2 / 5 / 16 / 3 / 2 / 1 / 2
Notes: See notes under Table 6.

Technology shares,
%
(N / C / H / W)

Marginal
technologies

35 / 0 / 63 / 2
44 / 0 / 54 / 2
34 / 0.3 / 64 / 2
29 / 3 / 65 / 2

N, C
N, C

5.8 Environmental effects


The model results show that applied production technologies vary widely. Market behaviour does not only affect prices and production quantities, but also applied production
technologies. Since different production technologies have different environmental impacts, model results indicate environmental effects of liberalisation and different market
conditions.
17

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

It is not meaningful to compare the production technology mix in the REF case with the
other cases, since the technologies used in the REF case does not reflect the actual production technologies that was used in the base year. Therefore, we are comparing the
COMP case with the STACK and STRA cases.
Prices are higher in the STRA case than the COMP case, which makes it possible to
profitably produce with higher-cost technologies in the STRA case. This could lead to an
increase of applied production technologies. Model results shows that this is indeed true
in all countries we find more applied technologies in the STRA case than in the full
competitive market situation.
However, there is no clear relation between which technologies that is used in the STRA
and COMP cases. The result depends on the market shares of the firms in the different
cases and how production technologies are distributed among firms. For example, in
France the share of nuclear production decreases and CHP and hydro production shares
increase when moving from a perfectly competitive market to a Cournot oligopoly market. In Germany, the share of nuclear production decreases, while the coal production
share increases. The STACK case either collapses with STRA, in case where the leader
does not have substantial market power, or leads to a mixed case between COMP and
STRA. Then STACK has the consumer advantage of a price lower than STRA and the
environmental advantage of a production lower than COMP.
Therefore, the results and country comparisons are very ambiguous. Higher electricity
prices do not necessarily lead to higher gains as electricity demand is decreasing. Nuclear technology is environmental friendly from the climate perspective as it does not
produce greenhouse gases like coal or oil plants. However, nuclear produces wastes that
harm the environment. Firms apply a specific technology primarily in the context of their
economic effectiveness and not because of the environmental hazard. Because of that,
we can detect that countries apply renewable technology primarily because of its technology and natural availability (like hydro in Finland, Denmark and Norway). The market situation does not necessarily bring an increase of environmental friendly technology. However, because a fully competitive market situation forces firms to choose market prices at their marginal production costs, low cost technologies are favoured that are
not necessarily environmental friendly.

6. Conclusions and Outlook


Model simulations of the European electricity market show that firm behaviour varies
substantially after a market liberalisation. In some countries, simulations show that
prices will go down irrespective of the resulting behaviour. However, in other countries,
competition is already so severe that prices can only rise in order to avoid bankruptcy of
power firms.
The main objective of liberalisation of the electricity market is to bring the prices down.
We find the lowest prices in the case with perfect competition (Finland, Germany). Alternatively, firms may also revert to strategic action, but dependent on the technology
18

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

production costs (Finland, Norway) and market power (France), it may or may not make
a big difference. It also matters to which extent consumers are sensitive to wholesale
market prices of electricity. A higher sensitivity may offset windfall profits under strategic action (France).
The liberalisation of the European electricity market leads to an application of low cost
technologies which are not necessarily environmental beneficial. However, in some
countries with a high share of renewable energy production technology (Norway, Denmark, Finland) strategic market behaviour increases the share of environmental friendly
applied production technology.

Acknowledgements
Funding from the Ministry of Science and Culture in Germany to build the model as well
as funding from the EU (contract number NNE5-2001-00519) for the EMELIE project is
gratefully appreciated. Any remaining errors are ours.

References
Amundsen, E. and Bergman, L. (2002) Will cross-ownership re-establish market power in the
Nordic power market? The Energy Journal 23(2):7395.
Andersson, B. (1997) Essays on the Swedish Electricity Market.
Dissertation, Stockholm School of Economics, Stockholm.
Bigano, A. and Proost, S. (2002) The opening of the European electricity market and environmental policy: does the degree of competition matter? Center for Economic Studies,
Katholieke Universiteit Leuven. Available from:
http://weber.ucsd.edu/~carsonvs/papers/602.pdf
Bower, J., Bunn, D. W. and C. Wattendrup (2001) A model-based analysis of strategic
consolidation in the German electricity industry, Energy Policy 29:9871005.
Bower, J. and Bunn, D. (1999) A Model-Based Comparison of Pool and Bilateral Market
Mechanisms for Electricity Trading, London.
Brubakk, L., Aaserud, M., Pellekaan, W. and von Ostvoorn, F. (1995) SEEM An Energy Demand Model for Western Europe. Report 95/24, Statistics
Norway, Oslo.
Cardell, J., Hitt, C. C. and Hogan, W. W. (1996) Market Power and Strategic Interaction in
Electricity Networks, Harvard Electricity Policy Group, Cambridge, Mass.
Dawson, Paul and Shuttleworth, Graham (1997) Efficient Transmission Pricing in Norway and
Sweden, Report for NVE, National Economic Research Associates, London.
Day, C. and Bunn D. (1999) Generation Asset Divestment in the England and Wales Electricity
market: A Computational Approach to Analysing Market Power, London.
Drake, F.-D., T. Ohler, et al (2000) Vom Preiskrieg zum Servicewettbewerb,
Energiewirtschaftliche Tagesfragen 5:286291.

19

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

European Commission (2003): Second Benchmarkreport of the implementation of the European


Electricity and Gasmarket, Brussels.
Ferris, M. and Pang, J. (1995) Engineering and Economic Applications of Complementarity
Problems. Wisconsin.
Ferris, M. and Sinaoiromsaran, K. (1998) Formulating and Solving Nonlinear Programs as Mixed
Complementary Problems. Wisconsin.
Fudenberg, Drew and Tirole, Jean (1992) Game Theory, 2nd print, MIT Press, Cambridge, Mass.
Green, Richard (1997) Electricity Transmission Pricing An Internal Comparison, Department for
Applied Economics and Fitzwilliam College, Cambridge.
Grohnheit, P.-E. and Larsen, H. V. (2001) Balmorel Data and Calibration.
Version 2.05. Available from: http://www.balmorel.com.
Helman, U., Hobbs, J.B., Cardell, T., Luong, T. and Wander, M.T. (1999) Modelling Strategic
Pricing in Bilateral and Poolco Electricity Markets: A Nash-Cournot Approach with Applications to US Eastern Interconnection, INFORMS national meeting, Philadelphia, Npv. 7-10
Jing- Yuan, W. and Smeers, Y. (1999) Spatial oligopolistic electricity models with Cournot Generators and regulated transmission prices. Operations Research 47(1):102112.
Kemfert, C. (1999) Liberalisation of the German Electricity Market - Strategies and Opportunities. Nota di Lavoro, 95-1999.
Kemfert, C. and Tol, R. S. J. (2000) The Liberalisation of the German Electricity Market Modelling an Oligopolistic Structure by a Computational Game Theoretic Model. Working
Paper 89/2000, Oldenburg.
Kemfert, C., Barbu, D., Kalshnikov, V. (2002): Economic Impact Assessment of the European
Electricity market, Working Paper, Oldenburg
Koopmans, C.C., te Velde, D.W., Groot, W. and Hendriks, J.H.A (1999) NEMO:
Netherlands Energy demand MOdel a top-down model based on bottom-up
information. CPB Research Memorandum (155), The Hague.
Lise, W., Kemfert, C. and Tol, R. S. J. (2003) Simulation of the Process From a Monopolistic To
a Liberalised Electricity Market in Germany. Submitted to Energy Economics.
McKinsey (2001) Paying a green premium, The McKinsey Quarterly, 2001 Number 3.
Murphy, F. H., H. D. Sherali, et al (1986) A mathematical programming approach for determining ologopolistic market equilibrium. Mathematical Programming 24:92106.
Newberry (2000) Privatisation, Restructuring, and Regulation of Network utilities. MIT Press.
Newberry (2002a) Regulatory Challenges to European Electricity Liberalisation. Cambridge.
Newberry (2002b) Problems of Liberalisation of the energy utilities. European Economic Review, Vol. 4.
Oschmann, V. (2000) Das Gesetz fr den Vorrang Erneuerbarer Energien. Energiewirtschaftliche
Tagesfragen 6:460464.
Prognos (2000) Energiereport III: Die lngerfristige Entwicklung der Energiemrkte im Zeichen
von Wettbewerb und Umwelt. Stuttgart, Schffer-Poeschel Verlag Stuttgart.
Pfaffenberger, W. and Haupt, U. (2001) Wettbewerb auf dem deutschen Strommarkt- Zwei Jahre
nach der Liberalisierung, 2. IEWT 2001, Strategien und Instrumente fr den Wettbewerb in
der Energiewirtschaft.
Rutherford, T. (1993) MILES, A Mixed Inequalitiy and non Linear Equation Solver.
Salant, S. W. and Shaffer, G. (1999) Unequal treatment of Identical Agents in Cournot Equilibrium. The American Economic Review 89(3):585604.

20

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Stern, J (1998) Competition and Liberalisation in the European Gas Market: A Diversity of Models, London.
Strombasiswissen (2002) Kooperationen unde Fusionen verndern die Stromlandschaft.
Available from: http://strombasiswissen.bei.t-online.de/SB115-04.htm
Tirole, Jean (1988) The Theory of Industrial Organisation. MIT Press, Cambridge, Massachusetts.
Verbndevereinbarung (2001) Verbndevereinbarung ber Kriterien zur Bestimmung von
Netznutzungsentgelten fr elektrische Energie und ber Prinzipien der Netznutzung.
Wietschel, M., Dreher, M. and Rentz, O. (2001) Grne Angebote in Deutschland: Stand und
Perspektiven; 2. IEWT 2001, Strategien und Instrumente fr den Wettbewerb in der
Energiewirtschaft.

21

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

7. Mathematical Description of the EMELIE Model


The computational game theoretic model EMELIE is characterised by following indices,
parameters and variables:
Indices:
- Firms f F
- Technologies i I

f
i
Parameters:
cvi
d0
p0

qmaxi,f

- Variable production costs for technology i


- Reference demand for electricity
- Reference price for electricity
- Price elasticity of electricity demand
- Maximum production capacity with technology i in firm f
- Electricity net transport losses

Variables:
p
cmf
sf
qi,f

- Demand price for electricity


- Marginal and average costs of electricity production of firm f
- Supply of electricity by firm f
- Production of electricity by firm f with technology i

EMELIE is a partial general equilibrium model of a liberalised electricity market with


multiple actors and is inspired by the theory of industrial organisation (Fudenberg and
Tirole, 1992; Tirole, 1988). On the supply side, electricity producing firms maximise
their profits. On the electricity demand side consumers maximise utility. In equilibrium,
prices p clear at national markets.
Note first that the market needs to be closed on the (consumer) demand side. This is
achieved by well-known inverse demand function:

p
sf = d 0

f F
p

0 p

(1)

Let us now consider the case with strategic interaction among firms at the supply side.
In this case, electricity producing firms f maximise their net profits. They do this by
choosing their strategies as represented by their supplies sf simultaneously by assuming
that other firms do the same. This is equivalent to maximising the profit firm f can obtain
by supplying to the grid. This profit is the difference between incomes from supplied
electricity minus the cost of production:

Maximise f ( S ) = ( p ( S ) c mf ) s f

(2)

where S = f F s f

(3)

where the dependence of the demand function p(.) on S constitutes strategic action,
while from equations (1,3), the explicit functional form of p(.) can be derived.
22

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

The first order conditions for optimality of strategically acting firms follow directly from
(1), by taking the partial derivatives with respect to sf:
f
f F : c mf = p 1 0 s f ,

(4)

The sign is used to denote the dual variable to this equation. The equation on the lefthand-side holds when firm f has a positive supply, a result which is well-known from the
Karush Kuhn Tucker conditions and is a typical characteristic of a mixed complementarity problem (MCP).
Equation (4) shows the equalisation between marginal costs and marginal income. The
individual market shares in equation (4) are conveniently determined by:

f F : f =

sf

gF

0 f

(5)

The marginal income in the case of strategic action is reduced by the factor market
share divided by price elasticity of demand. The market share represents a monopoly
mark-up.
Power supply by firm f to the grid can be generated with various technologies i, like nuclear, coal, lignite, gas, oil, hydro, and so on. We assume here a % electricity transport
loss, which is generally in the range of 05%.
f F : (1 ) qi , f = s f 0 c mf

(6)

iI

In order to fully define the model, we need to restrict firms marginal costs. For that purpose we have used variable cost data per technology, which do not differ per firm. Then,
a logical lower bound of marginal costs are these variable costs.

i, f ( I , F ) : civ c mf 0 qi , f qimax
,f

(7)

where the available technology is restricted by an upper bound too.


Model (1,47) is used to calculate the Nash equilibrium in the case with strategic action,
where market information is typically incomplete and we are dealing with the case of
imperfect markets. This case of quantity competition shall be referred to as STRA.
A model with perfect markets is established by replacing equation (4) by:
f F : c mf = p 0 s f

(4)

Furthermore, in the case of equation (4), equation (5) should be eliminated as well from
the model, as market share f is no longer a variable. Of course, the market shares now
follow exogenously from the model. Model (1,4,6,7) is used to calculate the competitive
equilibrium in the case without strategic action. This case is derived by reducing the demand function to an identity: p(.)=p. Here firms take market prices as given and we are
dealing with the case of perfect markets. This case of price competition shall be referred
to as COMP.
23

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

These model relations are written by the programming language GAMS, which decomposes the non-linear program as a mixed complementary problem (MCP). This is solved
by the non-linear MCP-solving algorithm MILES, which is a mixed inequality and nonlinear equation solver. Partially, MILES approximates linear sub-problems by Lemkes
algorithm and solves the non-linear program by the generalised Newton algorithm iteratively with a backtracking line search. An optimal solution is found by maximising regional profit conditions reciprocally under all considered constraints.
It is also useful to verify whether the initial prices and demand are viable. This is done in
the so-called REF case, where the production cost is minimised. This is expressed in
the following equation.
Minimise Cost ( qi , f ) = iI f F qi , f civ

(8)

Model (1,68) is run as a non-linear programming problem to establish the REF case.
Main outcomes of the model are regional prices, interregional trade flows and the optimal market shares of each electricity producer from which the regional concentration of
the industry can be calculated in terms of the Hirschmann-Herfindahl index (HHI). HHI
is a measure for (regional) competitiveness (see also Tirole, 1988, pages 221223). For
the industry as a whole the Hirschmann-Herfindahl index is calculated as follows:

s f ,r

HHI = 10000 rR
f F s f , r
rR f F

(9)

A fourth possible market situation is where one firm is a leader and moves first, while
the others are followers and move second in reaction to the chosen supply by the leader.
The followers behave just like the STRA case (while the fringe behaves as COMP). The
leader chooses its output knowing ultimately the outcome will lie on the followers reaction function (4). The followers reaction function is a constraint for the leader. The
leader chooses its output to maximise profit given this constraint.
The first order condition of the leader can then be written as:

c mf = p 1 f

dS
ds f

(10)

where dS/dsf =1 for the follower who behaves as in the STRA case, while dS/dsf > 1 for
the leader, which reduces the monopoly markup to the advantage of the leader. Hence,
the outcome of the Stackelberg behaviour will lie between COMP and STRA.
Derivative dS/dsf does no longer cancel out in the case of the Stackelberg equilibrium, as
the terms in the aggregate demand S = sf is built up from the supply of all firms, which
can be expressed implicitly via (4) in the supply of the leading firm.

24

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

Let us now simplify the model of f to two firms, where i is the leader and j is the follower.3 Then the reaction function sj(si) can be derived by rewriting (4):
( p c mj )

s j = si
p ( p c mj )

( p c mj )
p
= si
S = si + s j = si 1 +
; where j = p ( p c mj )
m

p
c
( j )
j

dS
p
dp dS 1
=
+ si c mj
dsi j
dS dsi j 2
dS
p
=
dsi j

si c mj dp
p j
1
= 2
2

j dS j si c mj

dp
dS

(11)

we know from the definition of inverse demand (1) that:


dp 1 p
=
dS S

substituting this in (11) this leads to:


p j
dS
= 2
dsi j + pc mj i

(12)

Finally substituting (12) in (10) and after rewriting the expression leads to:
(1 )2 + (1 )
f
f
f
f
c = p 1

(1 f )

m
f

p c mf
; where f =
; the Lerner index

(13)

Equation (13) is used as the FOC for the Stackelberg leader, while the followers are
competing in quantities (4), whereas the competitive fringe is a price taker (4).

To derive, mathematically, the FOC for the leader with n followers, is not an easy task, as
expressing the reaction functions of each follower leads to n equations, with n variables. This
can be aggregated to total demand, but the derivative w.r.t p is very complex. The present
approach nevertheless already shows the advantage of moving first.

25

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

8. Figures

Figure 1: Degree of Liberalisation of World Countries (Red: Fully liberalised; Blued:


Liberalisation underway)

30
25
20
15

in %

10
5
0
-5
-10
-15
-20
-25
AT

BE

DE

DK

ES

EU

FI

FR

% Change since 1/1999

GR

IT

LX

% Change since 07/2001

Figure 2 Percent Change of Consumer Electricity Price


26

IR

NL

PT

SW

UK

The European Electricity Market Does Liberalisation Bring Cheaper and Greener Electricity?

50

40

30

in %

20

10

-10

-20

-30
AT

BE

DE

DK

ES

EU

FI

FR

GR

% Change since 1/1999

IR

IT

LX

NL

PT

SW

UK

% Change since 07/2001

Figure 3:Percentage Change of Electricity Prices of Large Industries

140
120

in Euro /MWh

100
80
60
40
20
0
1

10

11

12

13

14

15

16

17

18

hour
Germany (E.On)

Austria (APCS)

UK

Figure 4: Difference of Average net Purchase and Sales Prices in 2002

27

19

20

21

22

23

24

Vous aimerez peut-être aussi