Title:
Asset development in the energy sector:
Optimal investment timing for a gas fired power station in the Netherlands
Master thesis Technical Business Administration Main subject: Financial Management
Author: D. Adema s1094386
Supervisors: Dr. W. Westerman (RuG) Ir. P. Spaans MBA (PwC)
University of Groningen
Faculty of Business Administration & Economics PricewaterhouseCoopers Advisory N.V.
Energy & Utilities
Summary
During my internship at PricewaterhouseCoopers Advisory N.V. (PwC) I wrote this master thesis. The department I took part of was Energy & Utilities, where I investigated the optimal investment timing for a gas fired power station in the Netherlands.
The reason why I choose this subject is as follows:
There have been many press releases from the large energy companies about investment plans for new power stations, but none of them has started actual construction. The issues are not the choice of capacity nor the location, but the fuel choice and the timing of
investments.
The fuel choice logically follows from forward curves: what will be the future price of coal, uranium or natural gas? To answer this question, extensive and specific market knowledge is a precondition. The logical approach to this question would be to predict the forward curves by the use of a model. Energy companies have much experience and knowledge in this field. The added value of a master thesis on this matter would not be high. Therefore I choose to concentrate on the timing issue.
The timing issue is not easy to estimate, because one has to rekcon with a lot of factors: strategic behaviour of competitors (it is not wishful to create over-capacity), prices of different fuels, development in electricity demand and drawing up the state of inventory of the current total generation park.
The thesis question therefore is:
Question: “What is the optimal investment timing for a gas fired power station in the Netherlands?”
The approach to tackle this question was to first make an inventory of the factors influencing the investment decision. Therefore, interviews have been conducted with energy sector experts within PwC.
The following factors came forward:
The (future) electricity prices Timing of investments
Financing Capacity choice
Fuelprices Location
Governmental regulation
At this point the critical investment factors were known, but it did not result in a direct approach to the investment timing issue. This is mainly because the factors: electricity price, fuel price and the timing of investments are related. (The wholesale electricity price are largely dependent on fuel- and CO2-emission costs.) The other factors have been
At that time however, the issue remained: how to tackle the research question? It has been chosen (of I have chosen) to approach the question by building a model, because the earlier mentioned factors in turn have relationships with other factors. A model can provide quantitative insight into the mutual dependencies.
The model has been constructed following the merit order system, a system that has been used in the Netherlands by the ‘Cooperating Energy Producing Companies’ (SEP) and which has been used in Britain. The idea is as follows: there is a continual electricity demand, which is called base load. During daytime, electricity demand increases and during night-time it decreases again. In order to generate exactly as much as is demanded, two types of power stations are needed: one covering base load demand and one covering peak load. Electricity cannot be stored and therefore generation always has to level
electricity demand exactly. This has rather large implications: besides providing for the constant level of demand (base load) one has to posses over power stations with the ability to be quickly heated up and become operational. The relative cheap nuclear- and coal fired power stations do not poses this quality. Therefore another type of power station is required: the gas fired power station.
Knowing this, the model has been constructed. Depending on the level of electricity demand, the power station with the lowest marginal costs is consecutively brought in to action to cover demand. The analogy is graphically exposed in the figure below.
The marginal costs of each power station can be estimated by calculating the operational efficiency; how much electricity is generated by how much fuel? This information can be obtained using the governmental-environmental annual reports; these reports are
mandatory to inform the government about the emissions of chemicals from the power stations.
There are 27 power stations with an environmental reporting obligation. All of them have granted a copy for this thesis. After a while, all the efficiencies of the power stations had been calculated and 26 of the power station could be included in the model. Information
concerning electricity demand was downloaded from EnergieNed.nl and Tennet.nl (the Dutch Transmission Support Organisation)
The course of electricity demand has been rather steady during the last 10 years; it increased with 2.5% on average over the ten years period. This steady increase of
electricity demand may change and therefore three different scenarios have been used in the model; low-, mid- and high growth in electricity demand.
The model could now be filled with the efficiencies of the power stations and with the three scenarios of electricity demand. The only things missing were the fuel prices and CO2-emission prices. (Because the electricity price of a nuclear power station is
determined only for a very small part by uranium prices, the nuclear power station of Borssele has not been included in the model.)
The future natural gas prices could be fed into the model by using natural gas futures; contracts for the future delivery of natural gas (unfortunately these futures trade with a
maximum expiration date of 3 years). Price information concerning coal and CO2
-emissions are not at everyone’s disposal. To settle this problem the trend in the historic prices has been extrapolated.
Next the model was filled with all relevant data. The only thing to do now was to
simulate the different years in the model and then calculate the expected electricity prices with and without bringing in the new theoretical power station.
Besides the three electricity demand scenarios, the model was filled with four capacity development scenarios I, II, III and IV. (Each of the scenarios I,II,III add more capacity and the fourth adds none). The influence on the future electricity prices of each capacity
development scenario and each electricity demand scenario has been compared. This results in 12 different combinations, each with different electricity prices for the next 15 years.
The next step was the compare the marginal costs of the new theoretical power station with the future electricity price (for each year). For the future electricity price the peak load price was taken, because a gas fired power station is a peak load power station. Now we want to know when the new theoretical power station has to be connected to the grid. The new theoretical power station has marginal costs of say Xi (i=1..15). The future electricity price, Yi had been calculated after including the new power station in the model (for each year i). The proceeds of the new power station will be Yi-Xi.
Because the new theoretical power station is a peak load power station, it will operate 5400 hours per year. For example: The difference between X1= € 40,- and Y1 = € 50 (per MWh) is the free cash flow of the new theoretical power station. It has been multiplied by 5400 to get annual cash flows. (Actually, O&M costs have to be withdrawn to get to real free cash flows.)
For each year and each combination of scenarios this has been executed. This led to 12 matrices with cash flows to the new power station. At times when these cash flows are higher, it is more attracting to have the new power station operational.
The cash flows in the twelve matrices have been set back to one present value. The discount rate was set equal to the overall discount rate in the electricity production sector using Bloombergƒ.
The present values have been calculated for the years 2006..2010. Subsequently, the cash flows of the next 12 years have been used to calculate the present value. Each time, the present value has been discounted for the start date; 2006.
This has been done for each combination of years and scenarios, the results are presented in the matrix below:
I II III IV Chance: 50% 35% 10% 5% 2006 low 25% € 42 € 9 € 3- € 103 mid 50% € 164 € 65 € 4- € 262 € 974 high 25% € 102 € 61 € 9 € 164 2007 low 25% € 52 € 13 € 2- € 13 mid 50% € 186 € 86 € 4 € 275 € 1.001 high 25% € 114 € 74 € 15 € 171 2008 low 25% € 59 € 17 € 2- € 112 mid 50% € 202 € 106 € 13 € 279 € 1.253 high 25% € 121 € 86 € 86 € 173 2009 low 25% € 66 € 21 € 21 € 0-mid 50% € 214 € 125 € 125 € 23 € 1.083 high 25% € 127 € 97 € 97 € 165 2010 low 25% € 72 € 25 € 25 € 117 mid 50% € 226 € 143 € 143 € 289 € 1.543 high 25% € 133 € 108 € 108 € 155 € 1.880 € 1.037 € 636 € 2.300
For example: connecting the theoretical power station to the grid in 2008, under the assumptions of mid-electricity demand growth and capacity development scenario III, resulted in a present value of 13 million euros. (The cash flows of the next 12 years have been discounted to 2008. And then 2008 cash flows have been discounted to 2006.) In the matrix, the present values of each year have been summarized. The same has been done for each capacity scenario. That way, the matrix gives insight to the expected proceeds of the new power station in each combination of scenarios.
The matrix makes two things clear: the highest proceeds will occur when the capacity increase will be zero (scenario IV). This scenario however is not likely; there would be shortages.
1stmodel outcome: New capacity has to be installed before the end of 2010.
In scenario I there will be just enough new capacity installed to fulfil electricity demand. This results in the second highest scenario. The summarized result of low/mid/high electricity demand growth is presented in the most right column. The highest returns occur in 2010.
2nd model outcome: Connecting a new power station to the grid in 2010 gives in the
highest returns.
The thesis question has been answered now; 2010 is the optimal investment timing. It is, however, interesting to compare the model results with “the-20%-reserve-margin” rule of thumb.
The next figure shows the capacity development over 30 years. The black line represents electricity demand (2.5% yearly increase) and the light blue line represents the case in which no new capacity will be installed. The dark blue line represents the ‘likely’ capacity scenario, scenario I.
The purple columns on the x-axis show the reserve margin; the margin between electricity demand and the line of the likely scenario. The grey line has been drawn through 20% reserve margin. The rule of thumb states: capacity should be added whenever the reserve margin falls beneath 20%.
0 5 10 15 20 25 30 1 9 9 0 1 9 9 5 2 0 0 0 2 0 0 5 2 0 1 0 2 0 1 5 2 0 2 0 T h o u s a n d s 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 R e s e r v e m argin P e a k d e m a n d s t a t u s q u o Likely s c e n a r i o
Based on the figure investment in new capacity is required, because there will be a shortage otherwise. This is concurrent with the model’s results.
Furthermore, it becomes clear that in 2008 the 20% margin will nearly be reached. Seen from the perspective of security of supply, it would be good to install new capacity in 2008. Again, the model gives similar results; the second highest present value occurs in 2008.
The data on which the figure above is based on, represents real capacity, in contrast to the model. The model only captures 26 large power stations. Nevertheless, the results of the model approach and the reserve margin match.
The thesis brought forward some interesting points:
The calculated electricity prices (by the model) are for a large part lower than the real electricity prices. This makes sense, because the model’s results are only based upon fuel prices, emission prices and efficiency. Operational and maintenance costs (and profit) have been left aside. The best fit between the calculated
electricity prices and the real electricity prices occurs when the first ones are being raised with 9 euros per MWh. In a publication of Energy Reseach Center (ECN) 9 euros per MWh is used as the margin for O&M-costs. This fact enhances the confidence in the model.
At the moment it is relatively cheaper to generate electricity using coal than by the use of natural gas. One disadvantage of coal, however, is the inherent pollution. A
coal fired power station emits on average approximately 800 kg of CO2per MWh,
a gas fired power station only 500 kg per MWh. At the moment, the cost of
emitting 1000 kg of CO2is around 15 euros. The model has been used to determine
the intersection: at which CO2-emission price will natural gas be a more attractive
fuel than coal (under the assumption of constant fuel prices)? The model indicates that this point will be reached at a CO2-emission price of 36 euros per ton. Again
this holds, at the end of April 2006 the CO2-emission prices were around 32 euros
per ton. At that time there were press releases pointing out that natural gas as a fuel was almost cheaper than coal.
A third point is the critical capacity border; the point of capacity where a subtraction of capacity (a power station) immediately leads to a change in the (wholesale) electricity price. According to the model results, this point is located at 17.334 MW. After adjusting this figure to the whole generation capacity of the Netherlands (only 26 power stations were included in the model) it becomes clear that the figure corresponds to the current level of generation capacity. In other words; if a power station of only 500 MW would be taken out of production, the electricity price would likely increase immediately. Recently (August 2006) there was a heat wave in the Netherlands, causing the cooling water to become too hot. Power stations could only generate electricity at a part of the total capacity. This had large effects on the electricity price. Although this situation is far more extreme than the withdrawal of 500 MW capacity, it can function as an indication that the generation capacity in the Netherlands is just enough to supply demand.
Table of Contents
CHAPTER 1: PROBLEM DEFINITION ... 1
1.1 RESEARCH DESIGN... 1 1.1.1 Execution ... 2 1.1.2 Key-investment drivers... 2 1.2 RESEARCH METHODS... 6 1.3 TIMETABLE... 7 1.4 CONTENTS THESIS... 7
CHAPTER 2: DATA ANALYSIS... 8
2.1 ELECTRICITY DEMAND... 8
DEMAND COMPARED TO CAPACITY... 9
2.2 ELECTRICITY PRICES... 12
2.2.1. Base load and Peak load... 13
2.2.2 Off peak- and peak prices ... 15
2.3.1 Electricity import and export (interconnection)... 16
2.3.2 Upcoming new investments in generation capacity ... 17
2.3 CORRELATION BETWEEN DEMAND,CAPACITY AND ELECTRICITY PRICES... 19
2.4 SPARKSPREAD... 20
2.5 CHAPTER CONCLUSION... 21
CHAPTER 3: COMPARISON OF THE DUTCH AND THE BRITISH ELECTRICITY MARKET . 22 3.1 BRITISH HISTORY... 22
3.1.1 Construction of the pool price ... 24
3.1.2 Complications of the pooling system ... 24
3.2 RESULTS OF DEREGULATION AND PRIVATIZATION... 25
3.3 COMPARING THE BRITISH WITH THE DUTCH SITUATION... 26
3.4 CHAPTER CONCLUSION... 26
CHAPTER 4: MODEL: INVESTMENT TIMING AND CAPACITY CHOICE ... 27
4.1 CHOOSING THE KIND OF MODEL... 27
4.2 SETTING UP THE MODEL... 29
4.3 SCENARIOS FOR CAPACITY INVESTMENTS WITHIN THE NEXT 5YEARS... 30
4.4 THE SHORT RUN MARGINAL ELECTRICITY COSTS CALCULATION MODEL (SRMECC-MODEL)... 32
4.5 MODEL RESULTS ASSESSED... 36
4.6 DEVELOPING THE DECISION METHOD... 38
4.7 RESULTS... 41
4.8 CHAPTER CONCLUSION... 43
5. CONCLUSIONS... 45
5.1 RECOMMENDATIONS... 46
LITERATURE... 49
APPENDIX 1 – EXISTING LARGE POWER STATIONS IN THE NETHERLANDS... 50
APPENDIX [2] EPPRICES COMPARED TO APXPRICES. ... 52
APPENDIX [3] SRMECC-MODEL ELECTRICITY PRICES PER MWH. ... 53
APPENDIX [4] SRMECC-MODEL PRICES DIVIDED OVER THE SCENARIOS... 55
APPENDIX [5] PRESENT VALUE CALCULATION OVER THE SCENARIOS. ... 57
Introduction
The energy sector and more specific the electricity generation companies are on the eve of a new investment wave. This is necessary, because if there would not be any investments in production capacity, there could be a shortage around 2008 in the Dutch market1,
according to Capgemini.
A number of energy companies are contemplating about new investments; this can be either in new plants or revising of older ones. Recently Electrabel showed its intention to build one new power station on the Maasvlakte near Rotterdam and to revamp two existing power stations in Flevoland. Appart from Electrabel, Nuon, Eneco and Delta have similar plans. At the moment, none of the companies has started construction yet. The companies have prepared the plans for the construction of new plants, but the final decision has not yet been taken.
Chapter 1: Problem definition
Without investments in new production capacity, it is expected that there will be a
shortage in the Netherlands around 20082. The ‘Financieel Dagblad (17/01/’06)’ claims that
if all current investment plans will be executed, total production capacity will increase with 30%.
Total production capacity in the Netherlands was around 18,000 MW in 2005. With an increase of 30%, this would become 24,000 MW.
There are however some uncertainties, for example the requests for building permits have been put in, but none of the companies have started building the intended plants. But also the amount of capacity needed is not known in advance and also is thus the investment value.
1.1 Research design
First, the factors for asset development in the energy sector are determined. The key-factors defined here are those key-factors that largely contribute to the success of an
investment in electricity generation. Second, these factors are analysed on their behaviour in the past. Using the historic behaviour of these factors, an assessment will be made about their future behaviour.
The key-factors are determined for a gas fired power station with a name plate capacity of 500 MW. An increase in production capacity of less than 500 MW, is relative small to the total capacity in the Netherlands and is therefore not representative.
1.1.1 Execution
To conduct the study, the research domain is restricted. The entire process starting with the idea until the construction of a power station takes a number of steps. Only a few of these steps will be included into this study.
The research sphere is on the meso level of the economy. The transition area between the micro and the meso level of the economy will be analysed to determine which investment factor is important in the decision making process for building a power station.
In figure 1, a detailed view is given of the research design. The factors below are presumed to be important in the investment process of a power station.
Figure 1, Research area.
Two factors will not be analysed in further: ‘geographic location’ and ‘regulation’ (the two factors in the red frame). The reason for this exclusion is the size of scoping for the thesis.
1.1.2 Key-investment drivers
The above introduced research area is being composed after first research and intuitively selecting the investment drivers. But are these really the key-investment drivers which are important and relevant for building new power stations? This question will be answered using two different approaches to warrant multiformity:
1. Interviews with energy sector experts from PricewaterhouseCoopers.
2. Using the findings of a CERA3report about energy investments world wide.
Factors 1 and 2 will be worked out further hereafter.
I Interviews
During the interviews the energy sector experts have been asked which elements they see as key-factor for optimal results for investments in a gas fired power station. Operational management is a factor in the success of a company. While answering the question
however, the assumption is made of a proper operational management and therefore it does not turn out as one of the key-factors.
The interviewees, all PwC, are:
Peter Spaans, Director Energy, Utilities & Mining. Formerly worked at Royal Dutch Shell.
Pjotr Schade, Advisor Energy, Utilities & Mining. Formerly worked at RWE and
Electrabel.
Paul Nillisen, Assistant Manager Transactions. Formerly worked at Nuon.
Frank de Lange, Director Valuation. Experience of valuing energy specific assets.
The responses on the interview question (key-factors for success in new electricity generation) are the following:
Electricity price
Financing
Fuel price/ natural gas price
Governmental regulation Timing/Capacity Location Sparkspread4= Thermal gas y Electricit
)
(
Price
)
(
Price
t
t
ηThermal= efficiency = (calorific value of the input fuel) / (energetic value of the electricity
output)
No new factors came forward during the interviews. The electricity price and the fuel price are covered by fuel/electricity modelling. Governmental regulation and location are treated further because of the size of scoping for this thesis.
4Spark spread = the difference between the market price of the generated electricity (in €/MWh) and the natural gas price, while keeping in mind the estimated efficiency of the power station (Electrabel, 2006).
II CERA about investments in energy
A CERA3(2005) report, The investments of the future: Toward a regulation ‘in concert’,
concerning the need for energy investments writes:
“Complicated investment risks demand effective risk management tools. Energy companies have to choose a mix between natural gas and electricity. The uncertainty is mainly about fuel prices, volumes, political environment and regulation. The best way to constrain risks is the possibility to enter in long-term contracts.
Because there is uncertainty about demand development of natural gas, new investments in security of supply are required. Concerning electricity production this mostly is the supply of natural gas as a fuel. Investments concentrate mainly on gas production but also on pipelines, LNG terminals and other storage facilities. (Recently Andris Piebalgs,
European Commissioner, proposed to store at least 2 months of gas for each EU member.) The industry primarily aims at new projects, long distance pipelines and LNG terminals. Considerations about the optimal capacity and timing of these projects will determine the success of them. To go successfully by these investments, one has to focus on the
fundamentals – the supply/demand balance. However, new risks in energy investments are: uncertainty about environmental protection, geopolitics of energy supply and perpetually changes in regulation. “
It is noted here, that CERA too refers to capacity and timing as being the most important issues for the success of investments. Based upon the interviews and the CERA report, for this study the key-investment factors are:
Electricity price
Fuel price, natural gas price Timing / capacity
All these factors are covered in the research design, and hereby the scope has shown to be legitimate.
The approach to the electricity market will be concurrent with the CERA proposal, on the basis of the supply/demand balance. The demand/supply balance is being assessed by comparing the electricity peak demand and the installed capacity, a figure of this is shown on page 19. (Furthermore, on the end of chapter 4, a prediction of the installed capacity and the prolonged peak demand is given.) Governmental regulation and location are not treated as relevant; refer to the preconditions on page 7.
Based on the supply/demand balance a model is going to be constructed to generate an (forecasted) electricity price. The difference between the costs of the new gas fired power station and the market electricity price are the proceeds of the new power station.
This leads to the main thesis question:
Question: “What is the optimal investment timing for a gas fired power station in the Netherlands?”
To facilitate this, the generic dependences of fuel, capacity, timing and price are going to be assessed. A model has to elucidate the way the variables are interconnected according to the electricity price. The choice of the connection with the electricity price is made because the electricity price is the basis of the revenue of a power station.
The next step is to assess the optimal time to connect the new power station to the national grid. The (financial) optimal time of connection is determined by the model. Before an actual connection can be made, one has to recon with building lead time and delay for building permits and operating permits.
The approximation of the timing issue is being handled using electricity future prices and regression curves for demand. The regression is done based upon historic data, which will not lead to an exact forecast, but it will give some notion of the investment impact. All this is done under different economic situations. During those situations different electricity demand curves are used.
This leads us to the following sub questions:
Sub questions
What is the importance of asset development timing regarding to power stations?
Which (statistical) relationships are there between capacity, price and demand?
What is the ‘critical capacity border5’ (determined through model simulations)?
Which time interval gives the best results for investments in a gas fired power
station?
5CCB; generation capacity can be added to a certain level and still have relative small effects on the electricity price. After this level/border the addition of new capacity will lead to significant changes in electricity prices.
Preconditions
The geographical location to build a power station has not been taken into
account;
Government regulation is out of the scope of this thesis;
The thesis handles the effects of a gas fired power station with a name plate capacity of 500 MW;
The determination of the optimal investment timing, the answer to the research
question, focuses within the next five years;
During the determination of the key-investment drivers correct/efficient
operational management is presumed.
1.2 Research methods
Research methods are practical means by which data can be gathered and analysed (Jonker & Pennink 2000). The methods are multiple and include the following:
Data is gathered on the internet site of EnergieNed, it uses that data to accomplish its yearly report ‘Energie in Nederland 2005’. TenneT, the Dutch TSO (transmission service operator) publishes a lot of electricity market data on its site. Especially the data of electricity load on the national grid will be used. This information is provided per 15 minutes and is called a PTE (per time unit). Using these rather extensive (35,000 rows of data per year) Excel exports, an Access database is constructed. This makes it easier to get specific information using SQL (Structured Query Language).
All the large electricity generation companies (Electrabel, Nuon, Essent, E.on and EPZ) have provided their governmental annual year reports. These reports provide information about emissions of CO2, NOxand SOx and the fuel- and chemical use per power station.
Using these reports the efficiencies and the CO2-emission rates of the power stations can
be calculated at normal production rates.
To acquire historic data, accounts have been provided by ENDEX European Energy Derivatives Exchange N.V. as well as one by the APX group. ENDEX is the Dutch platform where energy, electricity and natural gas, futures are being traded. The APX group (Amsterdam Power Exchange) is the platform where electricity is traded for next day’s use. Historic electricity- and natural gas price information of the day-ahead market can be downloaded from the APX website. Natural gas- and electricity future prices can be downloaded form the ENDEX website. At the site of EEX (European Energy
Exchange), a German trading platform, historic coal- and CO2-emission prices and coal
future prices can be downloaded. These are input variables in the model.
These sources, combined with the news portal energiea.nl, fulfil the data supply on which the analysis is based. Earlier academic energy research is retrieved by the use of the internet; a lot of publications can be found this way.
1.3 Timetable
This is the time planning for the thesis. Writing the report is a daily matter that is why the light blue timeline is not interrupted.
Figure 2, timetable
Reseach thesis Januari February March April May June
Asset Development weeknr. 1..6 7..9 9..12 13..17 18..21 22..26
1. Exploring the energy market. Formulation thesis question. 2. Writing the research proposal. 3. Start writing thesis.
4. Assess which data is needed. 5. Case study of the British market. 6. Data aquisition.
7. Data analysis.
8. Model construction and testing. 9. Modelling the scenario's. 10. Finish thesis.
The first two months have been used to do research and writing the thesis proposal, this time is market orange. After point 3, the necessary data is to be collected, this holds for the Dutch as well as for the British market. To know which data of the British market is
needed, point 4 is taken care of. After the data acquisition, it has to be analysed and made concurrent. The construction and testing of the model and the modelling analysis takes most of the total time. Finishing the thesis is mainly about moving from a draft version to a final version.
1.4 Contents thesis
In the next chapter the electricity market is explored and analysed. We start with the electricity demand curve and the available capacity curve. It will be clear new capacity has to be installed to satisfy future demand.
In chapter three the development of liberalisation and privatisation of the British electricity market is treated. Which lessons can be learned on the eve of privatising the energy sector and possibly unbundling the network- from the generation part of the energy companies.
In chapter four a dispatch model based upon the efficiencies and emissions of 26 large generating power stations in the Netherlands calculates the electricity price given a specific level of demand. For every level of demand, the situation is examined of
including (and excluding) a 500 MW CCGT6. The difference between the electricity prices
as a result of including the CCGT represents the effects of increased generating capacity. This finally leads to the final chapter containing the conclusion on the thesis question: ‘What is the optimal investment time for a gas fired power station in the Netherlands.’
6
Chapter 2: Data analysis
In this chapter we start presenting the electricity demand curve of the past 15 years. Then the available electricity generation curve is added. In the next step the demand and the generation curves are continued until 2020. This will make clear that new investments have to be made.
2.1 Electricity demand
The figure below shows the electricity demand in the Netherlands over the last 10 years. The straight black line is the regression line, it has a R2 of 0,9724. This means that
electricity demand has been relatively steady during this timeframe.
Figure 3, Electricity demand in the Netherlands.
80 85 90 95 100 105 110 115 120 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 G W h .
Data source: EnergieNed.nl
Figure 4 on the next page represents the graph of the total electricity demand in the Netherlands; this includes imports (interconnection) from abroad (most importantly Belgium, France and Germany).
The graph shows a relatively steady increase of the electricity demand over the past years. The average increase over a 10 years period of electricity demand is 2 .7%.
Capacity development
Now, let us look at the graph of the installed electricity generation capacity in the Netherlands.
Figure 4, Installed capacity.
1 0 0 0 0 1 2 0 0 0 1 4 0 0 0 1 6 0 0 0 1 8 0 0 0 2 0 0 0 0 2 2 0 0 0 1 9 9 0 1 9 9 5 2 0 0 0 2 0 0 5 M W
Data source: TenneT, EnergieNed.nl
The input data of this graph is founded on a spread sheet [appendix 1] containing all large production stations in the Netherlands. The list consists of all the large- and most of the small and mid-sized power stations. Interconnection capacity is included too.
Demand compared to capacity
In the next figure the demand curve and the capacity chart are compared. Using a rule of thumb, there always has to be 20% reserve capacity7.
Figure 5, peak demand compared to installed capacity.
0 5 10 15 20 25 1990 1995 2000 2005 M W T h o u s a n d s 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 R es e r v e ma r g in P ea k d e m an d g e n e r ation c a p a c ity
Data source: EnergieNed.nl, TenneT
The figure shows that there has always been more capacity available than strictly needed; the blue line lies higher than the black line. The difference between those two is the reserve margin. On the x-axis a bar chart of the reserve margin is given. The straight line through 0.20 shows the reserve margin needed by the rule of thumb. Between 1995 and 2000 there has been nearly 50% reserve margin!
There is a difference between the first demand curve and the demand curve used in the comparison. The difference between those two is the first one shows average demand during a year and the second one shows the peak demand over a year. The reason that we changed curves has rather important grounds. In the next section we discuss why peak demand is used instead of average demand.
Demand/production data
Electricity demand over the years runs very steadily. This however is not the case when we zoom in to a day to day basis. The total generation capacity has to fulfil electricity demand at any time. One very large complication of electricity is that is cannot be economically stored in large quantities8. Because it is not wise put on more power to the
national grid than there’s being pulled off, production and demand are on an aggregated level always in equilibrium. Very tiny unbalances cannot be prevented. The balance works like a push and pull mechanism; whenever a power producer supplies power to the grid this results in a push effect. Whenever electricity is taken off from the same grid, this results in a pull effect. The frequency on which the Dutch grid operates is 50 Hz. If the frequency descends to less than 49.90 Hz TenneT gives the power producers a signal to increase their output, if the frequency climbs above 50.10 Hz, TenneT signals the producers to decrease their output. If immediate measures are required, TenneT can regulate the excess electricity away with large resistors. Demand therefore leads to a production that is (except for unbalances) always in equilibrium.
Now we know that electricity cannot be stored, this implies that there always has to be enough capacity available to fulfil demand. Not fulfilling demand leads to rather
undesirable situations; traffic lights will not function any more, some industrial processes cannot be intermediately stopped etc, computer systems go down etc. So, it is very
important to secure enough supply to meet demand. For this reason, peak demand is used to compare the reserve margin.
Extending the trend
No let us see what happens when we extend the demand curve. Demand over the last 10 years increased at a steady pace of 2.7 % a year; this trend is extrapolated. The extended electricity demand curve is compared with the capacity curve. In this example, no new
8The newest development on this matter is the Redox Flow Battery. This battery doesn’t contain acid and lead, but uses a redox reaction between two different fluids. The problem is that these fluids do not contain a high energetic value; therefore large quantities of fluids have to be used in order to reach a reasonable amount of capacity. For example, a 100 MWh Redox Flow Battery needs two tuns the size of an oil tun at a refinery plant [1] (tun = oil storage tank at the port of Rotterdam).
capacity will be installed. The electricity generation park is rather old at the moment. In the coming years, some of the power stations will be mothballed (taken out of
production). The next comparison shows that new investments have to be made.
Figure 6, demand and capacity extrapolated.
0 5 10 15 20 25 1990 1995 2000 2005 2010 2015 2020 M W T h o u s a n d s -0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 Reserve margin Peak demand generation capacity
Figure 6 clearly shows that the current production park cannot fulfil demand over the next years. If no action will be taken, lights will literally go out in 2015! To keep a reserve margin of 20%, new capacity has to be installed before 2012. (Besides installing new capacity, it is also possible to overhaul existing generation capacity).
So the figure makes clear that capacity is needed. And if we stick with the 20% reserve capacity rule there has to be installed new generation capacity around 2008. In history, the decision of installing new production capacity was as simple as this, the analysis above. However, since the electricity market is liberalised things have changed. Now everyone can start his/her own electricity business. Therefore, electricity prices are no longer fixed by the government, but they are the result of demand and available generation capacity. In the next section, more attention is given to electricity prices. As we will see, electricity prices tend to fluctuate heavily. Because of this fluctuation, proceeds for an electricity generating company are all but certain. After liberalisation of the electricity market, risk has increased for the energy companies. Supplying demand is not the only concern of an energy company; they would like to make a profit too. The rather simple decision rule of 20% reserve capacity shows no insight to profit or revenue for a new power station. This is why a model will be constructed. Using the model, the expected cash flows can be calculated. Based upon the height of the cash flows a well founded decision can be made.
2.2 Electricity prices
Electricity prices are settled at the APX; the Amsterdam Power Exchange.
The APX is called the day-ahead market because all the electricity traded will be delivered the next day. Besides the above mentioned market there’s the unbalance market and the future market. Every larger electricity consumer has to nominate the electricity demand or supply for the next day. If it turns out a different amount of electricity is used, this will cause unbalances on the national grid. TenneT is the organisation that looks after the balance of supply and demand on the national grid. In times of high load, it signals the generators to turn down production and vice versa. This of course is not free of charge. The unbalance market settles the costs due to differences between the nominations and the real use of electricity.
All hours 0 20 40 60 80 100 120 2000 2001 2002 2003 2004 2005 2006 € / M W h
Figure 7 shows that APX electricity prices fluctuate heavily; fortunately one can trade electricity futures too. This way the proceeds of future electricity generation can be clicked in. The future market is traded on ENDEX in the Netherlands. Besides futures and
forwards trading, electricity contracts change hands bilateral too at ENDEX. This trade is called OTC (Over the Counter) trade.
There are three markets:
1) The day ahead market (APX) 2) The unbalance market 3) The future market
The reason that there are so many electricity trading platforms is that electricity cannot be stored in large quantities, because there is no efficient and cheap technology to support it. This practically means that all the electricity demanded has to be produced at the same time. The cheapest (short run marginal costs) ways to produce electricity is by operating nuclear and coal fired power stations.
Albeit this is the case, not all power stations in the Netherlands are nuclear and coal fired ones.
Figure 7, APX electricity prices on a daily basis
This again has several reasons; let’s start with the case of a nuclear power station.
First, there has been a lot of resistance against the risks of nuclear production due to the large scale accident at Tsjernobyl. Second, there is not yet any risk free option to stash nuclear waste (however the quantity of nuclear waste at the Borssele nuclear power station is limited to about 3.8 cubic meters per year [2]). And third, the initial construction costs of a nuclear power station are much higher than the costs of building a gas fired power station. For a coal fired power station, the initial cost argument holds too, combined with the building lead time. This is for a coal fired- as well as for a nuclear power station much longer than for a gas fired power station. The main issue however with a coal fired power station is the air pollution of it (CO2, NOx, SOxand heavy metals).
When comparing the CO2-emission of a gas fired9power station to a nuclear one with the
same capacity, a fully in use gas fired power station has a CO2-emission of 1,769,958 ton
and a nuclear power station of 1,183 ton. It is for the reader to decide what is most
harmful to the environment. When a gas fired power station would be interchanged for a nuclear one, the saving for CO2-emissions would be around 51 million Euros (CO2
-emission contract price 11 April ’06). The last comparison does not quite fit though, a gas fired power station would not be used to generate electricity during all hours of the day; a nuclear power station would. But for reasons of conception it is an interesting
comparison. A power station that would generate electricity all hours of the day would be a coal fired (or nuclear) power station. But for a coal fired power station the CO2-, SOx
-and NOx -emissions remain the problem. And last but not least, neither a nuclear- nor a
coal fired power station can be started up or turned off in short notice. Electricity demand can change very quickly and because there always has to be equilibrium on the national grid, this is a very import issue.
2.2.1. Base load and Peak load
Electricity demand differs from time to time during the day. The constant demand, the base load, can be produced using nuclear- and coal fired power stations. As electricity demand increases during the day, new power stations have to be connected. These power stations must possess the property of being able to start up and to be turned off at short notice. Gas fired power stations have these properties; they can be operational within 30 minutes. This is why gas fired power stations are used to cover peak demand. The electricity generated to cover peak demand is called peak load. Given the fact that a peak load power station doesn’t always operate, operational management is relatively more expensive compared to a base load power station. Because natural gas is a more expensive (but cleaner) fuel than coal, electricity prices rise when peak load power stations are brought into action.
This is why electricity is more expensive at peak hours (8.00 h.-23.00 h.) and cheaper during off peak hours (23.00 h. < off peak < 8.00 h.). Figure 8 represents the typical graph of daily electricity demand.
Figure 8, Base load dotted line, peak load continuing line.
To get to an equilibrium price at peak demand, the power station with smallest marginal costs can offer its electricity first. At times when peak demand increases, consecutive more expensive – measured over their short run marginal costs - power stations reach the point of being cost-effective as a consequence of the higher electricity prices. Figure 9 provides a graphical representation of this.
The widths of the columns represent the capacity of the power stations. The height of the columns represents the marginal costs. In this figure, power stations {1,2} cover base demand, plants {3,4,..,7} cover peak demand. The line through the top of the columns represents the supply curve.
2.2.2 Off peak- and peak prices
Figure 10 consists of the APX electricity prices of the recent years on a monthly basis. It shows that there are different electricity prices for different demand levels. Off peak prices (mostly during nights) are covered by relatively cheap nuclear and coal fired power stations, that is why they are lower. The fluctuation in peak prices is much larger than the fluctuation in off peak prices. The proceeds a typical peak load power station are therefore too (more fluctuating). During daytime (8.00 h – 23.00 h) peak prices have to be paid to get peak load.
Peak prices, black line (during peak load)
Average prices, blue line (all hours prices/base load)
Off peak prices, light blue line (during not-peak load hours)
0 20 40 60 80 100 120 140 160 2000 2001 2002 2003 2004 2005 2006 € / M W h .
Because in this thesis, the optimal investment timing for a gas fired power station will be calculated, it is important to know what the typical proceeds are. Based upon figure 9 a gas fired power station will only supply peak load. This is because it has higher marginal costs.
Besides generating electricity on home ground, it is possible to import electricity from abroad. All European countries are connected to surrounding countries, this is called interconnection. Especially France and Germany are large suppliers of electricity to the Netherlands. This is because Germany has many lignite fuelled power stations and France has many nuclear power stations. Besides interconnection, cogeneration delivers a share of total supply. Cogeneration is mostly used by industrial firms like a refinery plant. Besides electricity it uses the steam of the cogeneration plant.
To be complete with the generation possibilities, interconnection and cogeneration are treated briefly.
Figure 10, APX electricity prices
2.3.1 Electricity import and export (interconnection)
The Dutch market depends on electricity imports for 15-20%. The interconnection capacity between the surrounding countries is at the moment 3,350 MW. There are plans for new interconnection capacity from Norway and the U.K. at the size of 1,900 MW [3]. Although the efficiency of transporting electricity from the origin abroad to the consumer in the homeland deteriorates - due to the physical resistance of the cable - the interconnections have been used to the max. To get an idea of how much electricity is lost during transport, consider the next example.
Example: Electricity loss through interconnection cables
The resistance of a cable is given by the follow formula:
R = (ρ * l) / A
ρ= specific resistance (of the material of the cable) | 17.10-9 Ωm | in Ω
l= length | | in m
A= area of the sectional plane | π*r2 | in m2
Consider a high voltage cable (interconnection cable) compounded from 700 copper wires with a diameter of 0,8 mm and a length of 100 km. This cable would have a physical resistance of: R = (17.10-9* 100.000) / (4.10-4)2 * π*700 = 4.83 Ω.
Given a capacity of 400 MW and an electric potential of 150 kV, this implies a current of
I = P / U I = ( 400.106) / ( 150.103) = 2.67*103A.
This generates a loss of I2*R; in this example this would be (2670)2 * 4.83= 34 MW.
Compared to the 400 MW capacity this stands for a loss of 8.5%.
The maximum length of this cable therefore would be around 1,180 km.
Cogeneration
The Netherlands relatively has the biggest share in cogeneration, when comparing to other countries. The main benefit of a CHP10station is the higher efficiency. According to
a survey of PwC ’06 [4] after the 25 newest gas fired CCGT11power stations recently built
in Europe, an average efficiency of 0.58 came out. The efficiency of most CHP stations is above 0.75. The difference lies in the fact that a CHP also delivers heat, which is used either in an industrial process (when available) or for city heating. So the electric efficiency is combined with the thermal efficiency.
10CHP = Combined Heat and Power. 11CCGT = Combined Cycle Gas Turbine.
According to Kees den Blanken [5] another side effect of a CHP is that it has much lower carbon hydrates emissions. A 1,000 MW power station compared to ten 100 MW CHP stations generate just half of the CO2-emissions. There are however some considerations
not to build a lot of CHP stations immediately. Each CHP has to be connected to the national grid and each CHP has to be operated. The last two points defunct for a large scale power station, only one grid connection has to be made and there are some scale effects for operating the facility.
2.3.2 Upcoming new investments in generation capacity
In figure 6 we presented the case in which no new capacity would be added or
overhauled. This is not a likely scenario. If a shortage would really occur, then electricity prices will go ‘through the roof’. Besides the negative social implications of an electricity shortage, the high electricity prices would attract investments in new capacity.
Since 2004 there have been much news reports about intended new investment plans in new electricity production capacity. It is either about new power stations or renovation of older ones. A list of investment plans on new stations is presented below:
Renovation New Electrabel 5 2-3 Nuon 1 E.on 1 Essent 3 1 Intergen 1 Eneco 1 Delta 1
In appendix [1] all large production plants and intended new plants are presented. When all these power stations will be connected to the grid, and none of the existing stations will be mothballed, total production capacity in the Netherlands will increase reasonably. At this point, there will not yet be a discussion on the probability of the continuation of the plans (on page 32 this discussion is held). For now, let’s consider the increase in production capacity, assumed all plans will be carried out. Figure 9 has been extended with news reports from Energeia.nl, data from Tennet and information from EnergieNed. This results in figure 11.
10000 15000 20000 25000 30000 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 M W
Data source: EnergieNed, Energeia & Tennet.
From 2005 until 2009 no large scale project are going to be accomplished, as the graph shows. It is interesting to compare this graph with ICAP OTC-futures. ICAP is a large derivatives broker, handling derivatives not listed on an exchange (bilateral trade). When comparing the total generation capacity with the OTC-futures, a spike can be spotted at the start of 2007. Without jumping to conclusions, it seems that investors (in futures) bare in mind the addition of new capacity of around 2008/2009. Starting from 2009, total generation capacity increases. This corresponds with the descending prices of the
2007/2008 and 2009 futures (cal 07, cal 08 and cal 09). The assumption is that future prices run slightly ahead of the increase in capacity.
Figure 12, Future prices.
0 20 40 60 80 100 120 140 160 sep-06 Oct-06 nov-06 dec-06 jan-07 feb-07 Q4-06 Q1-07 Q2-07 Q3-07 Q4-07 Q1-08 Cal-07 Cal-08 Cal-09 € /M W h .
Data source: ENDEX Dutch power futures (14 August ’06)
2.3 Correlation between demand, capacity and electricity prices
At this point the main subjects to assess the correlation between the electricity price, the generation capacity and the demand of energy have been dealt with.
Intuitively, one would guess there should be a negative correlation between electricity prices and total production capacity. So, whenever the capacity increases (more than demand), electricity prices should go down. Reversed, the same thing would be anticipated.
To start, let’s look at a graph composed of electricity demand, APX prices and total capacity. The electricity prices have been corrected for the natural gas price. Electricity prices in January were € 107.90 on average, this peak has been left offside otherwise it would overshadow the other prices (calibration).
Figure 13, Demand, APX Prices & Capacity.
Data source: EnergieNed.nl, Energeia.nl, TenneT.nl, Essent, Nuon, Electrabel, E.on, EPZ, Intergen, Delta.
A correlation can only be drawn over two figures, so we have to adjust the generation capacity with the demand figures. The result of this action is the lower graph of figure 13. The correlation coefficient between the (natural gas adjusted) APX prices and the
(demand adjusted) capacity turns out to be 0.6! Totally unexpected it turns out to be a positive correlation, meaning that electricity prices rise whenever capacity is added. Fortunately, it is not a significant correlation.
It seems to be that electricity prices react heavier on other influences, some of them being the weather, interconnection and other (contingency) factors. Possibly APX prices do not react so much on small additions of capacity. Maybe the increases of capacity have to be larger to sort effect.
How to get past this obstacle? A model is introduced in chapter 4; it can track the results of i.e. these parameters.
13000 14000 15000 16000 17000 18000 19000 20000 21000 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 0 10 20 30 40 50 60 13000 14000 15000 16000 17000 18000 19000 20000 21000 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 0 10 20 30 40 50 60
Capacity MW. Average APX Prices.
€/MWh. MW.
Some side remarks to explain the unexpected behaviour of the electricity price and the unexpected correlation figure:
1. August10th 2003; there was for the first time since the heat wave of 1994 a code red12, this caused electricity prices to be higher on average.
2. The fuel price can be influencing the measurements. To correct this, the electricity price should be corrected for fuel prices (which has been done just for natural gas prices in the above figure).
3. The peak in January 1999 prices can party be explained by the rise in crude oil prices. The natural gas price depends on the oil price through the price of heating oil, in the previous quarter. The heating oil price is denoted on the site of Platts13,
therefore it is called the P-value.
2.4 Sparkspread
A very important variable for an energy company is the sparkspread, as used on page 4. The sparkspread is a derivative of the measurement for the gross margin of a gas fired power station. The formula for the sparkspread is Pelectricity(t) – Pgas(t) / ηthermal. [7] So the
spark spread shows the daily margin between the costs of fuels and the price fetched by the sale of electricity. The expression spark spread is only used for gas fired power stations, for a coal fired power station it is called the dark spread.
When comparing the theoretical sparkspread with the electricity prices traded on the APX, two intersecting lines occur. The area between the natural gas price line and the electricity price line represents a theoretical profit of a gas fired power station; the area beneath the spark spread line represents a theoretical loss of a gas fired power station.
Figure 14, Spark spread of a 51% efficient power station.
€ 0,00 € 10,00 € 20,00 € 30,00 € 40,00 € 50,00 € 60,00 € 70,00 € 80,00 € 90,00 200 2 Mar ch May July Sep tem ber Nov em ber 200 3 Mar ch May July Sep tem ber Nov em ber 200 4 Mar ch May July Sep tem ber Nov em ber 200 5 Mar ch May July Sep tem ber Nov em ber
Natural gas price * Efficiency. (Euro/MWh.) APX price (Euro/MWh.)
Data source: APX, Dte
12TenneT can signal either a code- green, orange or red. A code green means that there is enough capacity directly available. A code orange means that the capacity available within eight hours has reduced to 1,400 MW. A code red means that the reserve/regulation capacity available within 30 minutes has fallen below 700 MW.
2.5 Chapter conclusion
If no new capacity would be added, a shortage would occur around 2014. The 20% reserve margin showed that new investments have to be made. Before the electricity market was liberalised this rather simplistic approach was followed. At current times this approach is not enough. Electricity prices are not fixed by the government any more. They are now the result of supply and demand. This has a large impact; because investors mainly desire a profit, the new power station should only be build when the expected margins are high enough.
Because there is a difference in electricity demand during the day and electricity can not be stored in large quantities a variety of power stations have been installed. All these power stations have, depending on the fuels they use, different occupation schemes. This has to be considered in the timing choice of building a (gas fired) power station. Besides the difference in dedication, one should consider the expected spark spread for
investment purposes. Using futures and the bilateral contracts traded on the ENDEX an energy company can fix a spark spread for a predetermined time. (This is done by selling electricity futures and buying a natural gas delivery contract simultaneously.)
There didn’t seem to be any correlation between the changes in generation capacity and the electricity price, but this thesis persists in taking the chosen course. Electricity prices largely depend on daily changing variables like the dedication of power stations, but most importantly on the weather. Because the steam running through turbines has to be cooled to become water again and restart in the process, the temperature of the cooling water is an important factor. Because the weather cannot be controlled, it isn’t included as a decision variable in this thesis. The focus is kept on the dedication and the diversity of power stations.
In the next chapter the liberalisation and privatisation of the British market is treated. It is said the British market runs ahead of the Dutch market, so we look at the lessons to be learned and how the privatisation can work out in the Netherlands as a parallel to the British market. The construction of the pool price – the system marginal price (SMP) – of the next chapter gives good insight in how the model of chapter four is constructed. (The same analogy of figure 9 is followed.)
Chapter 3: Comparison of the Dutch and the British electricity market
While analysing the Dutch market, the objective is to make a statement what should be the best time to invest in new generation capacity. The analysis in this thesis is mainly based upon historic prices, production/demand, fuel- and CO2-emission prices. The
decision whether or not to invest should be based on a larger foundation. The electricity market in the Netherlands has been liberalised, but not yet privatised. The results of such a measure has been that the energy sector itself did very well (high profits) in Britain. Which parallels can be drawn and how will they effect the investment decision?
Besides that, the British pool system is a good explanation of how the model in the next chapter is constructed. The pool system and the idea behind the model are the same, but in the model there are no constraints for network flow capacities14, the latter ended up in
gaming behaviour of the larger generators in Britain as described further.
The British electricity situation is said to be a good comparison for the situation in the Netherlands. This is because of the fact that the Dutch and U.K. market have many similarities; both markets have the structure of a few large competitors, the biggest five companies control more than 80% of the market share. Both countries are being, or have been liberalised; the Dutch market from 2003 onwards, the U.K. market starting from 1990.
3.1 British history
From 1947 until 1990 the Central Electricity Generating Board (CEGB) operated all
generation and distribution facilities in England and Wales15. The Company was vertically
integrated and consisted of generation, transmission and distribution. The CEGB functioned as a vertically integrated statutory monopoly. The generated power was supplied to twelve state-owned regional distribution companies, called area boards that served their specific territories. The area board actions where tightly coordinated with the CEGB. “The resulting behaviour was little different from that to be expected from a single fully integrated public corporation”[9].
Two factors drove the change of regulation: First, overruns in construction costs of plants were very large and yet the CEGB was planning more constructions. Second, the electric power sector was performing poorly; productivity was way beneath the manufacturing benchmark. The publication of a white paper called “Privatizing Electricity” instigated the Electricity Act, which was adopted by Parliament in July 1989.
Starting in April 1990, the electric power sector was taken apart. The electric grid was transferred to National Grid Company (NGC). The generating capacity was split up over
14
Every gridline has a specific throughput capacity, for example 1000 MW. When there are three power stations of each 500 MW connected to one backbone gridline of 1000 MW, they can not all generate at maximum capacity.
15
three companies: PowerGen (52% of total capacity), National Power (33% of total capacity) and Nuclear Electric (15% of total capacity). The latter was later in 1996 dived further in two companies, British Energy and Magnox Electric.
The goal of privatisation was more competition and higher efficiency, the expected results of the free market system. It seems rather strange that the generation capacity was
divided in only three companies. “A failed attempt to privatise the nuclear stations had been the main motive for creating a company as large as National Power, in the hope that it would be large enough to absorb the risks of nuclear power. The stations had to be withdrawn from the sale in November 1989, and there was not enough time for significant changes to the restructuring plan.” (Joanne Evans & Richard Green)[11]
In 1991 National Power, PowerGen and National Grid were privatized. The area boards were privatized too, into 12 Regional Electricity Companies (RECs). All the shares of National Grid Company were held by the RECs, these kept their local monopolies. The RECs were accountable for metering, delivering and billing of electricity consumed in their area.
Starting from that time, all consumers with relatively small demands (less than 1 MW), had to buy electricity via their local REC at regulated tariffs. (Consumers above the threshold could choose their suppliers.) Despite the fact that only the top layer of the market was free to choose their own suppliers, there were a few entering generation companies and new CCGT technology enabled a constrain on prices, because this technology allowed new entrants to build small scale competitive generation capacity at relatively low construction costs and lead time and with high electric efficiency rates. On top of that, Britain had the dispose of relatively cheap natural gas from the North Sea. The first new entrant, Lakeland Power began operating its new 230 MW CCGT a year and a half later (July 1992).
At first, transmission of the Natural Grid Company and distribution by the RECs were treated as natural monopolies and subject to price cap regulation by the Office of
Electricity Regulation (Offer). Large users of electricity with demand levels in excess of 10 MW had to directly buy their electricity at the wholesale electricity market. They were free in their choices, even to buy their electricity abroad through an interconnected line. The intermediate companies, with demand levels between the 1 – and 10 MW could chose to get their power directly from one of the large generators or through contracts with one of the RECs (they were not locally bounded). The first steps of disintegration of the electric power sector had been taken.
The market had just opened up for the larger customers, although they were just about 5,000 in size, they accounted for around 30% of total demand. This situation has been going on until 1994; at that time the threshold for free choice in purchase was lowered to 100kW. Four years later, the threshold was abolished totally and market opening was a fact on a retail level, granting small customers to choose between their preferred suppliers. The RECs were obliged to resell power at the settled pool price.
3.1.1 Construction of the pool price
The CEGB was disintegrated and divided into three generating companies and the NGC. The dispatch of least costs and system reliability was taken care of by a new institution; the Electricity Pool of England and Wales. The institution was operated by NGC. All generators who sold directly to customers or with generating capacity over 50 MW were required to sell their electricity through the pool. All generating companies had to inform the NGC about their generation capacity and their bid prices for the next day in time frames of 30 minutes. The least cost dispatch is the merit order in which the bids take place. At some level total demand is covered by a series of bids. The last bid is the system marginal price (SMP). This bid price should resemble the marginal costs of providing for the demand. All companies called upon for their bid prices, receive not their bid price, but the SMP actually. At times with high levels of demand, and relatively low levels of
supply, the difference between the bid price from high efficiency generators and the SMP can get rather large. The pool purchase price is the SMP plus a capacity payment for capacity offered, but not called upon. This capacity payment should resemble the
difference between the short run marginal costs and the long run marginal costs. The final price for electricity is constructed of the pool selling price and a so called ‘uplift’. This uplift consists of a compensation for declared but unused capacity, out of merit running costs16, pool operating costs and transmission losses.
3.1.2 Complications of the pooling system
The previously described dispatch system, in combination with pool prices, uplift and the use of forward contracts to avoid the peaks and spikes in the wholesale electricity price, should constitute the foundation of a well functioning electricity market.
The purpose of the deregulated market was to increase productivity, enhance the free market system and lower electricity prices. In a sense this has occurred, but it can be described to deregulation or by contingency factors either.
Several problems occurred, for example limitations on the transmission capacity. This results in a sub-optimal application of the merit order system. Generators with higher bid prices are in such a situation called upon, receiving their bid price to make up for their short run marginal costs. On the other hand, some plants with bids lower than the SMP, weren’t called upon because of transmission limitations. These plants received the difference between their bids and the SMP. This results in a higher pool price and eventually in a higher market price. And as can be expected when there’s an angle, this system has been exploited by generators who calculated and predicted possible
transmission limitations. With just three large generators, Nuclear Energy with the nuclear power stations provided base load. The other two, provided the mid- and higher segment of the rank order of bid prices. Because the generation portfolios of the two were much alike, they could predict reasonably well the dedication of the other’s power
stations along with the other’s operating and marginal costs. To get higher prices for their
16
electricity, they upgraded the marginal costs of the power station that was likely to have the SMP, just before the next station in rank. This way they did not change the merit order, but were able to get the higher price when their assessment was right and the station they thought would have the SMP. This goes for the low bid - not called upon stations too, their incentive was to bid as closely as possible to zero, because they would get the difference between their bid and the SMP.
Another way to influence the pool selling price was to withdraw capacity as available. This has been done by PowerGen in 1991. Because of the capacity withdrawal the event of a shortage became more likely and the compensation for capacity offered, but not called upon was increased to compensate for the necessary investments in new capacity. On the very day the uplift was increased with the new tariffs, PowerGen declared the unavailable capacity as available again and that way receiving the larger tariffs.
3.2 Results of deregulation and privatization
There were actually some positive outcomes of the privatization and deregulation. “CEGB productivity gained an average of 3.5% each year; the three new formed companies had productivity gains of around 25% per year. Nuclear Power increased output with 50% in four years, while reducing the number of employees with 50% too. Before market opening Nuclear Power was operating on an average capacity of 59%, after restructure this
increased to 71% in 1995. Electricity prices sank with around 14% between 1990 and 1997.”[12] These factors seem to be all cheering in favour of the restructure, but let’s not jump to easy conclusions.
There is the issue of the lower electricity prices; this seems positive on first sight, but the lower prices fell together with decreases in the prices of the fuels; coal and natural gas. Second, with an increased productivity of 22%, increased market share for Nuclear Power and the layoff of employees, the decrease in electricity prices of 14% seems to be feeble. On the side of that, the RECs supplied 57% of local customers with demands in excess of 1 MW in 1990. In 1993 this figure had shrunk to 43%, while long distance load supplied by RECs increased from 4- to 16% [12]. All of this signs the market was working, but it contrasts unfavourable with the large profits these semi-regulated companies were making.
The semi-government companies, the RECs, that held all the shares in National Grid Company were allowed to charge rather high prices to foresee in large investments in the near future. The large investments never came and the RECs were making heavy profits. Even more obnoxious were the large executive bonuses that were paid.
