Journal of Energy Markets

Derek W. Bunn

London Business School

The Journal of Energy Markets completes its first volume with four papers highly representative of the topics, analytical methods and type of impact that this journal has sought to achieve. The year 2008 has been one characterized by extraordinary volatility in oil prices and the inevitable consequent knock-on effect of that volatility to wholesale gas prices. The valuation and trading of gas storage facilities has unsurprisingly become an increasingly intriguing topic with obvious risk management and investment implications.

In the first paper of this issue, "Valuation of a natural gas storage facility," Kjaer and Ronn develop a numerical algorithm based on dynamic programming that usefully provides insights into hedging operations based on gas futures contracts rather than on the more theoretically attractive but less practical spot strategy.

The use of real options to understand underinvestment is an active research topic, of concern to both regulators and industry participants. Optionality in the face of uncertainty can delay investment and this is a well-known driver of cyclical investment patterns in the energy industry. The second paper of the issue, "Price-cap regulation and investment behavior: how real options can explain underinvestment," by Nagel and Rammerstorfer, relates this real options approach to regulation and shows how price caps can further induce underinvestment. This topic is discussed in the context of the European transmission infrastructure, where price cap regulation, investor optionality and concerns about resource adequacy clearly interact in a crucial way.

In the third paper, "Electricity price forecasting with a new feature selection algorithm," by Keynia and Amjady, the important topic of daily price forecasting is addressed from the perspective of wavelet transforms and neural networks. Accurate short-term spot price forecasting in electricity markets is crucial in managing operational risk, and the challenges of using computational learning for this forecasting in order to achieve robust results are still very large for model builders. This paper is innovative in proposing a feature selection algorithm to help the specification.

Finally, the Markov chain Monte Carlo estimation process has attracted a lot of practical attention and in the fourth paper, "Markov chain Monte Carlo estimation of a multi-factor jump diffusion model for power prices," Green and Nossman develop a model that extends the well-known Lucia and Schwartz approach by using jumps and stochastic volatility within a Markov chain formulation. The comparative results for the Nordic market are interesting and will motivate further consideration of these types of models in practice.