A thesis submitted for the degree of MSc of Econometrics. University of Amsterdam, August 2011. sergej.o [at] googlemail.com

Abstract
The thesis deals with fitting, modeling and forecasting of the retail consumer demand for natural gas. This is important and has a potential for numerous applications: gas portfolio risk management, keeping the storage system in balance, using it for valuation of swing contracts, gas consumption planning and other. Retail gas demand is weather sensitive and therefore can be better explained, compared to the demand from industries or power plants. A relationship between demand for the natural gas, weather and prices is of key focus. Such a relationship is studied by different tests and models for the demand of the natural gas, spot prices and weather are proposed. Econometric approaches are used to prove that spot or forward gas prices do not effect gas demand in a short-term horizon. Two models for residential natural gas demand are proposed. These are used to forecast demand for natural gas of residential consumers. Two weather models are considered and used to simulate gas demand. A comparison and discussion is provided. Data are obtained from three energy companies in the Netherlands and Belgium.

Keywords (max 10): forecasting; electricity markets; time-series analysis
Published in: Journal of the Operational Research Society
Production / Publication year: 2006
Co-author 1: Conejo, Antonio J

Forecasting electricity prices in nowadays competitive electricity markets is a must for both producers and consumers because both need price estimates to develop their respective market bidding strategies. This paper proposes different transfer function models to predict electricity prices based on both past electricity prices and demands, and discuss the rationale to build them. The importance of electricity demand information is assessed. Appropriate metrics to appraise prediction quality are identified and used. Realistic and extensive simulations based on data from the PJM Interconnection for year 2003 are conducted. The models proposed are compared with naïve and other techniques.

Keywords: Electricity spot prices, ARMA models, Structural time series, Forecasting
Published in: -
Publication year: 2002
Co-Author 1: Jaroslava Hlouskova
Co-Author 2: Stephan Kossmeier
Co-Author 3: Michael Obersteiner

This paper studies the forecasting abilities of a battery of univariate models on hourly electricity spot prices, using data from the Leipzig Power Exchange. The speci cations studied include autoregressive models, autoregressive-moving average models and unobserved components models. The results show that speci cations where each hour of the day is modelled separately present uniformly better forecasting properties than speci cations for the whole time series, and that the inclusion of simple probabilistic processes for the arrival ofextreme price events can lead to improvements in the forecasting abilities of univariate models for electricity spot prices.

Keywords: forecasting, load,
Published in: Energy Risk
Publication year: 2002
Co-Author 1: Aram Sogomonian
Co-Author 2: Glen Swindle

This article presents an integrated approach to modelling temperature,electrical loads and power prices. This modelling framework will be used to analyse and manage the risks associated with serving variable quantities of load. The objective of the analysis is a better understanding and control of the earnings volatility associated with variable load risk.
The short-term variability of electrical loads is driven by temperature, season and time of day. Serving full-requirement load obligations involves assuming the uncertainty in the amount of power that will have to be delivered on a given day and hour. Forward power purchases can only hedge the quantity expected to be delivered, leaving the supplier with the risk associated with serving the deviations – whether positive or negative – from the expected load. This risk is termed variable load risk.
The financial risks associated with serving variable load obligations are compounded by the correlation between load and prices. High loads are often accompanied by high prices and low loads by low prices. If a load supplier is left short in a high-price environment or long when temperatures and loads are moderate, its earnings will suffer.
Here we introduce a new methodology for forecasting and jointly simulating temperatures and electrical loads. The mathematical analysis is presented in conceptual terms. The applications chosen to illustrate the methodology are load forecasting used to value full-requirement obligations and short-term load forecasts based on weather predictions. A later article will focus on simulating prices in the power markets and will present an integrated approach to managing variable load risk and reducing the earnings volatility of companies serving full-requirement load obligations.

Keywords: Forecasting, electricity markets,
Published in: Energy Trading
Publication year: 2002

The art of electricity price forecasting bears little resemblance to the utility practice it evolved from: trying to predict fuel costs. To predict prices, utilities and energy traders and wholesalers must resort to modeling electricity markets—no easy task. Price forecasting has become increasingly necessary and more complex for all market participants, as is evident from the variety of approaches being used today.

Keywords: forecasting
Published in: Global energy business
Publication year: 2002

Winning in competitive electricity markets takes a fair amount of educated guesswork. Energy marketers cannot be certain that their future delivery of power at the price specified in a long-term contract will earn them a profit—because supply, demand, and the going rate may differ from expectations at the time the contract was signed. As a result, energy traders are only as good as the load forecasts they use.

Keywords: forecasting, risk management
Published in: Global Energy Business
Publication year: 2002
Co-Author 1: Richard Hooke


In competitive markets, all customers must be pursued, but all customers are not alike. Some bring more risk than reward to a marketer’s overall portfolio. Profitability demands that marketers price these hidden risks appropriately.

Keywords: Forecasting
Published in: Global Energy Business
Publication year: 2002

Competitive markets put a premium on information, and that information can become scarce when it's owned by market participants. A case in point is load forecasts, which utilities have perfected over many years of providing power to their service areas. Suddenly, market participants are active in areas outside of their service areas, or new entities are buying and selling power between regions, and no one has a clue what load will look like.

Keywords: Electricity, Volatility, Regime-switching, Structural Models
Published in:
Publication year: 2003
Co-author 1: Nektaria Karakatsani

This is a review paper documenting the main issues and recent research on modeling and forecasting electricity prices. The special market microstructure of electricity is described, as an explanation of the extraordinary stochastic properties of electricity price time series. The research literature deriving from the application of models adapted from financial assets, for both spot and forward prices, is reviewed and criticised. Final emphasis is placed upon the virtues of computationally intensive structural modeling.

Keywords: Electricity markets, forecasting, market clearing
price, time series analysis.
Published in: IEEE TRANSACTIONS ON POWER SYSTEMS
Publication year: 2002
Co-author 1: Javier Contreras
Co-author 2: Antonio J. Conejo
Co-author 3: Rosario Espinola

In the framework of competitive electricity markets, power producers and consumers need accurate price forecasting tools. Price forecasts embody crucial information for producers and consumers when planning bidding strategies in order to maximize their benefits and utilities, respectively. This paper provides two highly accurate yet efficient price forecasting tools based on time series analysis: dynamic regression and transfer function models. These techniques are explained and checked against each other. Results and discussions from real-world case studies based on the electricity markets of mainland Spain and California are presented.