Modeling the UK Electricity Market Using Quantile Regression: Scenario analysis of non-linear sensitivities to fundamental variables
MetadataShow full item record
This paper develops fundamental quantile regression models for the UK electricity price in each trading period. The sample covers half hourly data from 2005 to 2012. From our analysis we are able to show how the sensitivity towards different fundamental factors changes across quantiles and time of day. In the UK the supply of electricity is to a large extent generated from coal and gas plants, thus the price of gas and coal, as well as the carbon emission price, are included as fundamental factors in our model. We also include the electricity price lagged by one day, demand and margin forecasts. We find that the sensitivities vary across the price distribution. For instance, at lower prices the sensitivity towards coal increase, while at higher quantiles the sensitivity towards gas increase. Our findings also suggest that the sensitivity to fundamental factors exhibit intraday variation. For example, during low demand periods, like nighttime, today’s price is heavily influenced by yesterday’s price. During high demand periods, like midday and evening, changes in margin forecasts are more important. We have performed a scenario analysis based on our regression models revealing the actual impact on the electricity price when the values of the different fundamental variables vary within the range of their minimum and maximum values from our data set. Our findings are useful for risk managers, producers, consumers, buyers, traders and regulators, as they provide a measure of the price risk tied to different fundamental factors at different times and price levels.