Prices Across the Markets: How the German Coal‐Based Electricity Price May Affect the Swedish Nuclear/Hydro‐Based System

Abstract

The paper elaborates on the possible correlations between the electricity prices in two neighbouring regions. The question is of interest since a large part of the public and some politicians feel that the European CO₂ emission trading system ought not to affect the hydro‐ and nuclear‐based Swedish electricity market. The analysis builds upon empirical observations and simple econometric testing. Our results highlight some of connections between the Swedish and German electricity prices and we qualitatively discuss these results. Furthermore, we restate the obvious fact that the investment climate in Sweden must improve if the wish is to avoid having coal‐based electricity production on the margin.

αMats Nilsson, Luleå University of Technology and Vattenfall R&D, and Kristian Gustafsson, Vattenfall R&D. Mats Nilsson is the corresponding author: . E‐mail: [email protected], Södertäljevägen 1A, 64532 Strängnäs, Sweden. The opinions expressed in this article do not necessarily reflect the opinions of the affiliations but rest with the authors. The authors wish to thank Robert Lundmark for support and ideas of improvement. The usual disclaimer applies.

Keywords:

• Electricity
• regional markets
• policy

Notes

αMats Nilsson, Luleå University of Technology and Vattenfall R&D, and Kristian Gustafsson, Vattenfall R&D. Mats Nilsson is the corresponding author: . E‐mail: [email protected], Södertäljevägen 1A, 64532 Strängnäs, Sweden. The opinions expressed in this article do not necessarily reflect the opinions of the affiliations but rest with the authors. The authors wish to thank Robert Lundmark for support and ideas of improvement. The usual disclaimer applies.

1. Included in our definition of the Nordic market are Denmark, Finland, Norway and Sweden. Iceland is not interconnected with the other countries in the region. The use of the term ‘Scandinavia’ is inappropriate, as it only encompasses Denmark, Norway and Sweden.

2. See, e.g. Bergman (2005) or Damsgaard and Green (2005) for a comprehensive discussion concerning the development of the Nordic market.

3. Press release 30 August 2006, Swedish government. ‘Mona Sahlin kommenterar nedläggningshotet mot Utansjö bruk’

4. The structural break that may be envisaged in the figure is with all probability too large to have been created by this acquisition. See further, Nilsson (2005) for more information on this.

5. The dry winter meant that the energy produced by hydro in the Nordic system was far below that which is expected a ‘normal’ year. Von der Fehr et al. (2005) describes the dry winter and how the market reacted to this in some detail.

6. The consequences of the Green certificate system are not all clear. Bergman and Radetzki (2003) estimates ex ante a cost of 3 öre/kWh for the consumer. This is partially confirmed in Nilsson och Sundqvist 2006. On the other hand, the low variable cost of renewable electricity production should put a short‐run downward pressure on the electricity price.

7. http://abcnews.go.com/International/wireStory?id = 1461996

8. See, for example, Hjalmarsson (1996).

9. ECON (2004b) estimates that the probability of a similar situation occurring again is 1 in 200.

10. To study whether we have a structural break we have used a chi‐test, see for example Gujarati, (1992).

We first estimate the parameters in the following equation:

Then we divide the data set 20030101, and two equations are estimated. The null hypothesis is that these parameters should not differ. F‐value = 22.733274>F‐criteriont = 2.232079. Thus the null of no structural break is rejected.

11. The exact numbers are Y = −0.0796+179,1*time and Y = −0.1314+276.23*time. We urge that great caution is used when interpreting these numbers.

12. Sometimes in the literature these two are denoted the ‘add‐on’ and the ‘work‐on’ rate.

13. This simplification is problematic as we will have both problems with identification (is the change in the variable caused by a movement of supply or demand (or both), and with endogeneity as it is not clear to us that quantity causes price and not vice versa. Having said that, this simplified econometric study may give some quantitative evidence to a folk theorem circulating in the Nordic market in the last 5–6 years. See Gujarati (1992) for some further elaborations of these problems.

14. Due to auto‐correlation the Prais‐winstens method was used.

15. For the sake of clarity we will use this expression throughout to avoid the confusion with a confirmed causality, as this type of test simply indicates causality.

16. Thanks to Robert Lundmark for help with econometric work.

17. One way of dealing with this would have been to clean out these effects from the price series. As the results of the Granger causality are at best indicative, we have abstained from this econometric exercise.