|
Article Excerpt
Volatile fuel prices affect both the cost and price of electricity in a liberalized market. Generators with the price-setting technology will face less risk to their profit margins than those with costs that are not correlated with price, even if those costs are not volatile. Emissions permit prices may respond to relative fuel prices, further increasing volatility. This paper simulates the impact of this on generators' profits, comparing an emissions trading scheme and a carbon tax against predictions for the UK in 2020. The carbon tax reduces the volatility faced by nuclear generators, but raises that faced by fossil fuel stations. Optimal portfolios would contain a higher proportion of nuclear plant if a carbon tax was adopted.
1. INTRODUCTION
Emissions trading schemes and carbon taxes work by raising the cost of producing electricity from fossil fuels. The more polluting the technology, the more its variable cost increases. This should provide an incentive to shift generation towards low-carbon technologies. We know that taxes and permit schemes should have equivalent results in a world of certainty, but that they perform differently in an uncertain world. This paper considers their different impact on the profit risk faced by electricity generators with a choice of technologies.
The standard analysis of investment choices in electricity uses a cost-minimizing approach. Investors do not choose to build new power stations solely on the basis of their expected costs, however. Companies wish to make profits, and to avoid excessive risks. Roques et al. (2006a) show that a probabilistic analysis is needed to give the full picture of the expected return on an investment, and its distribution. One particular issue discussed by Roques et al. (2006b) is that if the cost of gas and carbon are correlated with the price of electricity, the profit margin of a gas-fired generator can be less risky than either its costs or its revenues, considered in isolation. The profit margin of a nuclear generator may be much more risky than that of the gas-fired station, since its costs will not be as correlated with the price of electricity.
This paper asks how the choice between a carbon tax and an emissions permit scheme (based on auctions) affects the risk of investing in gas-fired or in nuclear generation. With a carbon tax, the variable cost of gas-fired generation is raised by a fixed amount, which will normally feed through into electricity prices. With carbon permits, the price of emissions depends upon market conditions. In the specific case of the electricity industry, the key factor is the relationship between coal and gas prices. Taking a slightly Euro- and electricity-centric view, the price of carbon will be set so that the electricity industry within the EU reduces its emissions to the level that is needed in order that the demand for emissions permits equals the supply. Imagine that the main choice in the short run is between running coal- and gas-fired plants, (1) and that fuel prices are such that most coal-fired stations would be cheaper than gas-fired, in the absence of the ETS. In this case, this implies that the price of carbon has to rise until enough coal-fired output has been displaced by gas generation. The price of carbon is then at the level which equalizes the running cost of some gas-fired and coal-fired stations (Newbery, 2006). A rise in the price of gas will tend to raise the price of permits. The higher gas price thus has both a direct effect on the price of electricity, and an indirect effect via the cost of emissions.
The aim of this paper is to discover whether this magnified relationship between gas and power prices has a significant impact on the relative risk of different kinds of generator, compared to the alternative of a carbon tax that was not linked to the level of fuel prices. The paper follows the same basic approach as the papers already cited, but uses a more detailed model of the relationship between input prices and the wholesale price of electricity, that of Evans and Green (2005). This allows us to calculate the expected profitability of each type of power station, taking into account the way that its operating pattern will depend upon its variable cost relative to other stations.
The paper finds that choosing a carbon tax instead of carbon trading would raise the standard deviation of gas-fired generators' annual profits by nearly 40%, would double the standard deviation of coal-fired profits, and would reduce the standard deviation of nuclear profits by almost 40%. With emissions trading, the price of carbon varies to offset movements in relative coal and gas prices, reducing the risks faced by fossil-fuelled generators. A switch to a carbon tax removes this insurance. Since carbon prices are correlated with the gas prices that typically set the price of power, emissions trading makes the electricity price more volatile than a carbon tax would, and hence raises the volatility of nuclear stations' profits, since their costs are assumed to be fixed.
These changes raise the proportion of nuclear plant in the portfolios that optimize a generator's risk and return. While the impact depends on the level of risk aversion, adopting a carbon tax could raise the optimal proportion of nuclear plant from none to 16%. Most of the rest of the portfolios consists of gas-fired plant, although at high levels of risk aversion, some of this is replaced by coal generation. The paper also shows that hedging the output price of nuclear stations with long-term sales contracts would allow a much higher proportion of nuclear plant--up to 54% with carbon trading, and 64% with a carbon tax. While the focus of this paper is on nuclear plant, renewable generators with fixed costs face similar issues.
The next section briefly discusses the background to the Emissions Trading Scheme and the economics of a liberalized electricity industry. Section 3 outlines the supply function model used to determine the relationships between fuel and carbon prices and generator profits. Section 4 presents the data used, drawn from the DTI Energy Review (2006) and Supergen FutureNET scenarios (Elders et al., 2008). Section 5 presents the results for single plants, and section 6 discusses optimal portfolios containing a mix of plants. Section 7 concludes.
2. THE ELECTRICITY INDUSTRY AND THE EMISSIONS TRADING SCHEME
Traditionally, the electricity industry has largely been vertically integrated. Large companies that combined generation and transmission might sell power to smaller distribution utilities, but this was usually done via contracts or tariffs, rather than with any kind of market mechanism. Following Chile, England and Wales, and Norway, many countries have now adopted wholesale markets for electricity, with competition between generators. While the details vary across countries, the key elements are that generation has been split from transmission, that entry into generation is largely deregulated, and that generators compete to sell their output, through a centralized market, bilateral contracts, or both.
This has changed the way in which companies need to think about investment. Traditionally, investment plans were made with the objective of minimizing the expected cost of meeting the forecast level of demand. There could be a trade-off between capital costs and fuel costs--plant that was expected to run for most of the time could incur high capital costs in return for lower fuel costs (the standard example being nuclear power) while still being competitive against plant with lower capital costs but high fuel costs. For plant that was not expected to run for much of the time, it would not be worth incurring the high capital costs. Most newly-built plant was expected to run nearly continuously on base load, however, and so valid cost comparisons could be made on the basis of the expected cost per kWh generated at a standardized, high, load factor. The option with the lowest expected levelized cost would normally be the front-runner for investment. The classical investment appraisal appeared to take little notice of risk or uncertainty.
In a market-based system, companies invest to earn profits. Their first criterion will not be to minimize the expected cost of meeting demand, but to maximize the expected difference between their costs and their revenues. In a fully competitive market, in which the company was not able to influence the prices a plant received, this would come down to minimizing its expected cost, as with the pre-liberalization approach. This equivalence between the outcome of a perfect competitive market and a perfect social planner is, after all, one...
|
