3 Presentations

• 
  04:00 PM - 04:30 PM

  Nord Pool Day-Ahead Power Prices: Forecasting Switches between Volatility Regimes based on Fundamental Information

  • Ronald Huisman, presenter, Erasmus School of Economic
  • Sjur Westgaard, NTNU

  Price uncertainty in power markets varies over time. Price volatility is high when the market operates at its limits. Variation in the level of volatility is therefore not stochastic as fundamental information about the situation in the market might forecast volatility changes. Our goal is to examine power price levels and volatility in relation with fundamentals. We apply regime switching models to govern the time variation in volatility by modeling the transition probabilities as a function of fundamentals. We show that fundamentals explain a large part of price variation and that changes in volatility can be forecasted by supply and demand variables. This helps power producers and distribution companies with optimizing their nominations, with pricing and with measuring risk.

• 
  04:30 PM - 05:00 PM

  Demand-at-Risk Forecasting in Case of Seasonality, Extremes and Volatility Clustering

  • Verena Hagspiel, presenter, University of Lausanne

  We consider the problem of managing high levels of demand in presence of considerable uncertainty in the form of stochastic demand volatility, nonstationarity and extremes.
  We suggest a two steps model which allows the estimation of Demand-at-Risk.
  The first step filters data from stochastic volatility, the second step models extremes of demands in presence of nonstationarity.
  The method is illustrated on daily energy consumption data in three different regions of Norway over the past three years. The data contains stylized features of daily demand series such as seasonality patterns and dependence on external variables.

• 
  05:00 PM - 05:30 PM

  Switching Options in Peak Power Plants: Empirical Evidence

  • Stein-Erik Fleten, presenter, Norwegian University of Science and Technology
  • Alois Pichler, Norwegian University of Science and Technology
  • Carl Ullrich, Virginia Tech

  We analyze the real options to shutdown, startup and abandon peak power plants. We assume that the plants’ status for a given year is either operating, in standby or retired. The analysis is made using data for 1,121 individual power plants for the period 2001–2009, a total of 8,189 plant-year observations. We estimate the irreversible costs of switching by structural estimation of a real options model. The results indicate that plant managers have decision heuristics that take uncertainty into account in a manner consistent with real options theory.