Read more about 'Risky Power'

Risky Power: Choice of technology, security of supply, and market power in power markets

Blackouts are generally perceived to constitute an important risk to the workings of modern societies. To counter this, societies need to secure the availability of generation capacity. The way this is done has been substantially altered, as power markets have been liberalised, partly privatised and re-regulated. Increased reliance on renewable energy, use of market-based prices, and decentralised commercially-based decisions on investment in generation capacities have added new stochastic elements to electricity systems.

This project explores the interaction between risky supply, stochastic demand and the organisation of power markets and how these combine to determine the risk to supply security and adequacy – the risk of a blackout.

The first aim of the project is to analyse the optimal mix of technologies for power generation, when the costs and benefits related to the security of supply are taken into account. The second aim is to analyse the (short run) effects of market power on the choice of generation technology and to correctly determine the degree of market power when demand and some supply are random. The third aim is to assess the importance of market structure and conduct for investment in new capacity. In particular, we wish to assess the effect of different vertical structures of generating companies.

This research project is financed by the Danish Social Sciences Research Council with a grant of DKK 4,5 million (approximately EUR 600,000) and by Copenhagen Business School.

The research team consists of six researchers with strong research complementarities. The competences include knowledge of real options, investment in infrastructures, vertical structures of electricity markets, security of supply, intermittent power technologies and market power in power markets. Thus, as a collective, the group is ideally positioned to deal with the research questions. Below we summarise individual competences with a focus on those that deal with the project:

Fridrik M. Baldursson (Professor of Economics, University of Iceland) is an expert on real options and on power markets.
Anette Boom (Associate Professor of Economics, Copenhagen Business School) is an expert on investment in generation capacity, supply adequacy, and vertical market structures.
Stefan Bühler (Assistant Professor, University of Zurich) is an expert on infrastructures and vertical market structures.
Stine Grenaa Jensen (Scientist, Risø National Laboratory) is an expert on mathematical modelling and statistical analysis of the energy sector including non-renewables.
Poul Erik Morthorst (Senior Research Specialist, Risø National Laboratory) is an expert on economic and technical assessment of renewable energy technologies and on regulation.
Peter Møllgaard (Professor of Law & Economics, Copenhagen Business School) is an expert on market power and on industrial organisation in relation to competition policy and regulation. Project leader Peter Møllgaard has a strong experience in research management through involvement in e.g. two EU funded projects and through his position as head of CBS’ Department of Economics.

In addition, we plan to hire two PhD students to deal with two self-contained research projects that are embedded in the overall project. Description of these projects are found below.

MAIN PROJECT DESCRIPTION: OVERALL AND UNIFYING IDEAS

This project deals with the mix of generation technologies, the level of generation capacities, and security of supply in power markets. Risk enters at several levels. Demand for power is stochastic – but so is supply – and hence the risk of a blackout is real. Some generation technologies have a more intermittent nature than others: Energy produced by wind mills is fed into the transmission systems more randomly than power produced by traditional fuels. Thus, the mix of technologies affects the security of supply and the need to invest in reserve capacity. As power markets have been liberalised, decisions to invest in capacity have been decentralised and are now partly or wholly at the discretion of commercial companies. Their incentives to invest in generation capacities depend on market prices. How well markets work, depends on how they are organised and regulated.

Relevance: why is it important?

Blackouts, such as those that happened in 2003, are generally perceived to constitute an important risk to the workings of modern societies. To counter this risk, governments need to ensure the availability of generation capacity. The way this is done has been substantially altered as power markets have been liberalised, partly privatised and re-regulated. (IEA,2005).

Increased reliance on renewable energy, use of market-based prices, and decentralised commercially-based decisions on investment in generation capacities have added new stochastic elements to the electricity systems. Around the world, governments and market participants struggle to get a grip on supply security in the changing environment.

Wind power is a subsidised alternative to traditional electricity generation that by now constitutes a significant fraction of power supply, but although wind mills provide inexpensive electricity some of the time, it is difficult to use for securing supply. In fact, the existence of such intermittent supply influences the profitability and hence availability of generation capacities that are fuelled traditionally (oil, coal, and gas) and are typically used to secure supply. Wind power increases the necessity of having alternatives ready to secure supply, but at the same time it decreases commercial incentives to invest in such capacities. (Morthorst, 2004; IEA, 2006).

The first purpose of the project is to analyse the optimal mix of technologies for power generation, when the costs and benefits related to the security of supply are also taken into account. We take a market-based view of security of supply and extend the sparse existing literature (see Egenhofer et al., 2004; Joskow and Tirole, 2004).

The introduction of electricity markets implies that prices become increasingly stochastic. In the Nordic countries, electricity is traded on an hourly basis on the Nord Pool power exchange. Based on competitive bidding, market prices change from hour to hour. This in turn affects generators’ choice of technologies. In the presence of start-up costs, a traditional power plant will only start up, if the market price exceeds a certain benchmark price in excess of the marginal costs. The plant will stay open until the price falls under a lower benchmark price. This follows from basic real options theory.

The first implication of the real options approach is that the mix of technologies in operation at any point in time is determined by the market – although in a non-obvious way, since the owner may possess different technologies with different risk-profiles, essentially involving trade offs between the size of start-up costs against marginal costs. Furthermore, it is possible that the owner has market power, so the market price is not perceived as given, but rather something that the generator can influence. Thus, a second purpose of the project is to analyse the (short run) effects of market power on the choice of power generation technology.

A second and related implication of the real options approach to technology choice is that even in the absence of market power, the competitive price of the marginal plant will (at least at times) be above the marginal costs that could be estimated from deterministic data. A third result of the project should be to correctly assess the “risk-adjusted” marginal costs, and hence to determine the correct degree of market power.

The liberalisation of power markets not only meant the introduction of competition rather than regulation of vertically integrated monopolies, but also decisions about the vertical structure of the industry: Should the generator be allowed to operate the grid and own distributors? The Danish history is illustrative of this dilemma. In early re-regulation, liberalisation foresaw vertical separation of generators and distributors. Nevertheless, recently Danish generators have been allowed to merge with distributors to form vertically integrated companies.

It is an open question, which of these vertical structures gives the better incentive to invest in capacities and, thus, in supply adequacy. Buehler et al (2004), Boom (2002) and Boom and Buehler (2005) conclude that vertically separated generators will invest less in generation, but that all vertical structures will ensure supply adequacy. Yet, this hinges on the way blackouts are modelled and, thus, a fourth aim of the project is to study, which vertical structure that is best for society when blackouts occur in equilibrium.

Research questions

Overall the common theme to our research questions is the interaction between risky supply, stochastic demand and the organisation of power markets.

The first subproject analyses the optimal mix of technologies for power generation when costs and benefits related to the security of supply are taken into account. The subproject will investigate how far market based options for supply security can be pursued as a substitute for regulatory options.

The second subproject takes technologies and capacities as given and analyses the (short run) effects of market power on real-options based choices of power generation technology. The allocation between technologies at any given hour of trading will be determined using a real options approach and correct risk-adjusted marginal costs will be calculated.

The third subproject takes up supply adequacy and aims at explaining, which vertical structure will reduce the risk of inadequate supply the most and whether this is also the structure that serves society the best.

Theoretical foundations and methodology

In the tradition of new industrial organization, we develop and extend relevant theories and possibly test these using empirical techniques or else find relevant cases on which the theories have a bearing. For each subproject we first develop theories to fit the research problems that we identify in power markets; we will collect relevant data; and we will test the theories using quantative methods. As the relevant theories, the testable hypotheses and the analytical strategies are to some extent unique to each subproject, we relegate the discussion of these (and the synergies) to these. The practical feasibility of the project is also discussed at each subproject.

Embedded PhD project 1: Market-based security of supply

Security of supply, i.e. the ability of existing capacity to meet actual load (Joskow, 2003, 2005) is costly, so there is a trade-off between avoiding blackouts and paying for reserve capacity to be ready at any time. Essentially this trade-off is known from insurance markets. The purpose of this integrated PhD project is to identify, to which degree these risks are marketable and how this might be done. The project will be carried out by a PhD student in collaboration with the senior researchers in the team and supervised by Anette Boom (main supervisor) and Poul Erik Morthorst (co-supervisor).

There are a number of different ways, in which the balance between demand and supply of electricity may be maintained at any given point in time. These include reliability contracts and capacity markets on the supply side (see e.g. Egenhofer et al., 2004) and making some demand price responsive (see Hirst, 2002; Stridbæk, 2003). Presently, such market-based options for balancing supply and demand are not priced in the market and the degree, to which markets can price such “public goods” is an open question.

Consumers may play an increasing role in the marketability of the risks. The point of departure is that the typical time-invariant retail price paid for power includes two components: a payment for the electricity and a risk premium that protects consumers from the volatility of market prices by allowing them to buy unlimited amounts at prices that are fixed beforehand (see Hirst, 2002, Boom and Buehler, 2005). Policy-makers intend to make electricity demand more elastic and this entails transferring some of the stochastic nature of wholesale prices to retail prices at least for some consumers. This will increase security of supply to the extent that such large consumers react to price spikes that may be early warnings of supply disruptions.

In addition, there are suggestions that some large customers may agree to be cut off in order to reduce demand when blackouts are imminent. This could for example be a large freezing storage facility that can be disconnected at no or low cost. Essentially, a market for extra short term capacity and short term “disconnectability” could arise providing the TSO with the least expensive way of dealing with shortages or imbalances.

This PhD project first identifies the different instruments available to market electricity supply risks through a thorough literature review. It seeks to identify possible obstacles to the provision of “security services” such as their public good nature and free riding. It also aims at identifying, if similar insurance products can be provided through forward trading and if forward prices send the right signals to market participants.

It then seeks to quantify the relevant price and quantity risks using the project’s dataset. In doing this there will be a strong synergy with the real options approach. In essence the empirical part seeks to determine how the risk premium in the spread between wholesale and retail prices should be determined.

Embedded PhD project 2: A real-options approach to determining power prices

In the analysis above we have assumed that the stochastic processes were exogenous to the companies. A complicating factor arises when, as is the case with market power, the company may actually influence the pricing. A purpose of this integrated PhD project is to analyse the (short run) effects of market power on real-options based choices of power generation technology. This means that the loop is closed: stochastic prices may affect the dispatch of technologies in the short run, but the choice of technologies may also affect the stochastic price process.

In standard real options theory, the decision-maker faces an exogenous stochastic process and deals in the best possible way with this through a standard S,s policy. When the stochastic process becomes endogenous (meaning that the decision-maker may influence it), then the standard tools of real options need to be adapted. In such a strategic context, a real option to start up or close down a plant may allows the option holder to manipulate the distribution of returns and, hence, depends on both expected future payoffs and the variance of those payoffs (Nickerson and Sullivan, 2006). This problem is harder to solve.

The way this endogenous process relates to power markets could be the withholding of short-run capacity that would have been dispatched by a non-strategic company or in the longer run the under-investment in generation capacity. Both types of behaviour would tend to increase prices and increase the risk of blackouts. Such effects have been identified by Joskow and Kahn (2002) in relation to California’s electricity crisis (see also Joskow, 2001; Olesen and Sundahl, 2006).

This PhD project first identifies the relevant literature on strategic real options decisions and provides a thorough literature review. It then seeks to set up a model of strategic real options decisions in the power market that may be solved preferably analytically or else using numerical methods.

The project will be carried out by a PhD. student in collaboration with the senior researchers in the team and supervised by Peter Møllgaard (main supervisor) and Fridrik Baldursson (co-supervisor).

Sidst opdateret af Bente Faurby 06.03.2009