Modelling Correlation in Carbon and Energy Markets

Abstract

The paper examines correlations between daily returns of month-ahead baseload electricity, fuel input and carbon emission allowance (EU-ETS) prices for Great Britain. The perspective of a CCGT plant operator is assumed, producing baseload electricity with natural gas and emission allowances and selling output forward in the month-ahead market. Price correlation between power, natural gas and emission allowances as well as their dynamic behaviour is essential for the extent to which cashflows from CCGT plants are self-hedged. Switching between input fuels with different carbon intensities is taken as the fundamental driver of this correlation. Relative marginal power generation costs are used to construct carbon price regimes during which no switching takes place. The regimes are then used as explanatory variables in a dynamic conditional correlation model. Using daily observations of month-ahead prices from April 2005 to August 2010, the results suggest that extreme weather, high commodity market volatility and seasons have no effect on correlation. However, there is evidence of significant price decoupling during periods of extreme relative carbon, coal and natural gas prices.