Editor: Sjur Westgaard
Published: 26 Mar 2010
Papers in this issue
by Per Bjarte Solibakke
by Jonas Andersson, Jostein Lillestøl
by Margaret Armstrong, Alain Galli
by Anders Løland, Xeni K. Dimakos
by Virginie Dordonnat, Siem Jan Koopman, Marius Ooms
Norwegian University of Science and Technology
This special issue contains 5 papers that were presented at the seminar "Price and Decision Support Modelling in Electricity Markets" at the Department of Industrial Economics and Technology Management of the Norwegian University of Science and Technology. The purpose of the seminar was to stimulate the ongoing international dialogue among academics, practitioners and policymakers with mutual interests in energy markets and energy project investments. This was the first energy seminar organized with financial support from the ongoing FINERGY and ELDEV projects. These projects are financed by Agder Energi, Statkraft, Troms Kraft, Trønder Energi, the Research Council of Norway and the Adolf Øien Foundation. The workshop provided the opportunity for researchers and practitioners to discuss theoretical and empirical issues regarding the empirical modeling of energy markets, the pricing of energy derivatives, investment valuation of energy companies as well as real projects, and other decision support analysis within the energy sector. The papers in this issue involve electricity load and price modeling, market structure and risk modeling of energy markets. The data covers the Nordic, French and German electricity markets.
In the first paper, Andersson and Lillestøl model and forecast the electricity consumption of South Norway (the “NO1 area”) in the Nord Pool market using functional analysis of variance and functional regression models. Hourly data from 2004 to 2008 are employed. The former appears to be useful in studying the different seasonalities of electricity consumption; the latter is useful to improve univariate forecasts of hourly consumption.
Another paper modeling electricity loads is that of Dordonnat et al. In their study they look at French national data for the period 1997–2007. The original hourly time series is transformed into a daily vector of national demands. They investigate the different components that determine the load and the evolution of these components over time, with an emphasis on the intra-daily pattern. They employ a dynamic factor model with smoothness restrictions on the intra-daily load pattern within a linear Gaussian state-space framework. Their new modeling framework provides satisfactory insights for modeling and forecasting load patterns in France.
The French market is also studied by Armstrong and Galli. In their paper they review the structure of this market since its liberalization in 2001. In particular, they discuss the day-ahead auction market and the market for virtual power plant (VPP) options. Holders of VPP options have the right (but no obligation) to access electric power at a predetermined price per megawatt-hour, in 30-minute slices 24 hours a day, seven days a week, during a given delivery period. In the paper they show that the structure of the French market allows holders of aVPP option to exercise strategies of selling power on the exchange when the day-ahead price is above theVPP strike price and, conversely, of buying power on the exchange when the day-ahead price is below the strike price. By documenting the strike prices of VPP options that were active at different times and by examining the structure of day-ahead prices they demonstrate that participants actually follow this strategy. The authors also show that the presence and absence of peaks coincides with the strike price of active VPP contracts. The modeling of electricity prices is studied in Løland and Dimakos’s paper. This paper investigates the determinants of relative electricity price differences between the South Norway (“NO1”) area and the Nord Pool system price (whole area). The NO1 area is dominated by hydropower and will periodically experience a large deviation from the Nord Pool system price. Using price data from January 2001 to June 2008, the authors test the effect of explanatory variables such as seasonal factors, water reservoir levels, capacity variables and flow. Using a generalized additive model, they find that the NO1 price is below the system price when the NO1 water reservoir level is high relative to normal levels, and when the export capacity is limited.
In the last paper, Solibakke discusses risk management issues for energy companies. He first discusses corporate risk management at a general level and then moves on to discuss proper risk measures and models for estimating univariate and multivariate positions. He performs volatility, correlation and value-at-risk estimation of both the Nord Pool market and the Phelix market using several statistical techniques including generalized autoregressive conditional heteroskedasticity models, stochastic volatility models and extreme value theory.
The Trondheim winter energy seminar continued a series of seminars and conferences with topics related to decision models for energy markets.As many of the topics described and analyzed in this issue can (and will) be extended in various directions, we intend to follow up with another similar seminar during the winter of 2010.
Search the archive
Subscribe to gain full access to The Journal of Energy Markets and its valuable archive.
Call for papers
Submit your work and we can offer you:
Please contact email@example.com for more information.
Updating your subscription status
Risk iPad and iPhone Apps