CO2 and fuel switching in the power sector how econometrics can help policy making

Alexander Antonyuk, Gas, Coal and Power Division and Bertrand Magné, OECD Environment Directorate

The switching of fuels in power generation is the topic of much analysis in the field of climate policy. Fuel switching, especially from coal to gas, is an effective short- to medium-term measure used to abate carbon dioxide (CO2) emissions in electrity generation. The extent of its potential remains uncertain to policy makers. Econometrics based on market data can reveal the possible magnitude of fuel switching as a result of price changes, including those induced by a price on CO2 emissions. The method discussed here, and illustrated for the United States and Spain, can help to establish realistic short-term CO2 emissions reduction goals in competititive electricity markets.

The role of fuel switching in lowering CO2 emissions from electricity production

The decarbonisation of the power sector requires the implementation of regulatory measures and changes in business strategies. Setting the envisaged energy revolution in motion requires, among other things, a carbon-pricing policy to discourage the use of CO2-emissions-intensive fuels and generation technologies (IEA, 2010).1 The third phase of the European Union Emissions Trading Scheme (EU ETS) starting in 2013, together with the possible emergence of new carbon market mechanisms and carbon taxes (e.g., in Australia, Chile, some Chinese provinces, Japan, Korea, Mexico, New Zealand, and some US states), could trigger significant changes in the operation of electricity generators. In the short term, switching across fossil fuels, from coal or oil to gas in particular, may be an important transition measure to lower CO2 emissions from electricity generation.

Fuel switching in thermal power plants has been identified as one of the main CO2 emissions-mitigation options in the short term, and deserves careful assessment when designing policy. The extent to which reductions can be expected in the near future from coal-to-gas switching, for instance, is useful information for policy makers who set the overall cap on emissions.2

1. Dedicated incentives or price and research and development (R&D) mechanisms to support the deployment of low-carbon technologies (including fossil-fuel plants with CCS, nuclear plants and renewables-based electricity) need be implemented in parallel.

2. Fuel switching is meant here as substitution of fuels which is done using existing generation capacity, and is different from long-term, gradual changes of capacity and generation mix (e.g., phasing out oil-fired capacity and generation).

In the presence of a price on CO2 emissions, fuel switching will be the first-business response to reduce operating costs and CO2 emissions. It will also contribute to the reduction of other local pollutants (e.g. particulate matter), thus reconciling both utilities' interests and environmental concerns. For these reasons, understanding and characterising fuel-switching behaviour is crucial. Arguably, these mechanisms can be best observed in competitive electricity markets, where generators are meant to optimise operations to minimise cost.

This paper investigates the prospects for market-based power-fuel switching, looking at historical fuel use and price data and using relevant statistical methods (see Hicks, 1932; McFadden, 1963; and Morishima, 1967). The econometric methodology, described in full after the Conclusions, measures the responsiveness of fuel use in electricity generation to changes in international coal, gas and oil markets - in economics terms, it estimates fuel price "elasticities" of fuel demand. These estimates can then be used to forecast degrees of fuel switching in electricity generation. Although it is applied here to the power sector, the method can be used for other sectors where fuels can be substituted.

While the examples given below are mainly for illustration, they are based on real data and preliminary conclusions on fuel switching can be derived for the United States and Spain. The examples of these two countries are used in different ways: there is comprehensive data for the US electricity sector, which permits testing of the methodology; there is, however, no CO2 pricing applied across the United States. Spain, by contrast, is part of the European Emissions Trading System and the analysis can be used to demonstrate the possible fuel-switching effect of a CO2 price variation.

The relationship between fuel cost and fuel use

In order to identify the role of fuel switching among emissions-mitigation mechanisms, and the relationship between fuel switching and carbon prices, we introduce a proven method of quantification.

Since emission costs are added to fuel and other costs in calculating the price of generating a unit of electricity, the question of how carbon price affects fuel use can be answered through an investigation of how relative generation costs, based on different fuels, influence short-run switching between them. An estimate of this relationship allows us to forecast fuel switching driven by a change in carbon price: Change in CO2 price ^ changes in fuel generation costs ^ changes in fuel inputs (fuel switching)

The first calculation of the sequence is easily done using standard emission factors for each fuel and the carbon price observed on the market for EU allowances. The second calculation, which requires knowing the relationship between fuel costs and fuel use, is provided by the model we describe below.

The switching between coal and natural gas in the US power sector provides us with an illustration of how fuel use is affected by relative fuel prices. Although there is currently no carbon price in place at the federal level in the United States, the core of our analysis is an estimation of changes in fuel use in response to price changes.

Figure 1

Relationship between relative prices of coal and gas and their shares in thermal generation, United States

Average monthly fuels shares and gas price

Average monthly fuels shares and gas price

un 100% CD

0% 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Source: IEA statistics.

Notes: the shares are percentages of total generation from combustible fuels (gas and coal) and add up to roughly 95% each month, the rest being generation from oil and combustible renewables; the averages are for years 2003-2009 excepting 2005 and 2008, which show extreme fluctuations.

Figure 1 shows a clear correlation between short-term seasonal changes in fuel prices and fuel choice (note that the coal price does not demonstrate great seasonal fluctuation; the seasonal nature of natural gas fluctuations, however, creates seasonal variation in the price of coal relative to gas). These data, coupled with the fact that the US power sector has been partly liberalised, indicate a possible link between relative fuel shares and underlying fuel prices; for the purpose of policy making the important question is the magnitude of such a relationship. The model and measures used here are described in the Methodology section; they essentially rely on the notion of cross-price elasticity: that is, a percentage change in demand of one fuel divided by the percentage change in price of another, alternative fuel.

Empirical estimates: The United States and Spain

The US Energy Information Administration (EIA) provides monthly generation and fuel-cost data separately for utilities and independent power producers (IPPs), showing significant differences in capacity and generation mixes for each and allowing separate estimates. Table 1 shows results for utilities only.

Table 1

Cross-price elasticities in US utilities

Table 1

Cross-price elasticities in US utilities

Price of Demand of^--^





Was this article helpful?

0 0

Post a comment