Heavy tailed modelling of electricity prices and estimation of extremal correlation with the market

Specialeforsvar: Jakob Juel Kirketerp Nielsen

Titel:  Heavy tailed modelling of electricity prices and estimation of extremal correlation with the market

Abstract: The increase in energy prices in Europe since the Corona-virus pandemic as well as the Russian invasion of Ukraine has had a huge impact on worldwide financial markets. These energy price increases have made it more expensive for firms to produce goods, thus making the modelling of electricity prices essential for a firms ability to reduce costs of production. A central notion in heavy tailed theory is the concept of regular variation. Establishing some main properties of regular variation allows for a descriptive summary of heavy tailed modelling. Classical extreme value theory is used to model the heavy tailed behaviour of the energy prices. Other exploratory tools for heavy tailed time series are introduced and discussed. These include estimation procedures of the index of regular variation, which determined the heaviness of the tail of a distribution and estimation of the extremal index. The regular variation property is essential in construction the quantitative tool, the extremogram, introduced by A. Davis and T. Mikosch in [7]. This concept allows us to create an extremal analogue to the classical autocorrelation function. To create reliable confidence bands, a modified bootstrap method is used. We attempt to analyse the correlation be- tween large energy price increases and the German stock index, the DAX. The extremal dependence structure between the DAX index and electricity prices are analysed, to determine how extremal events affects the German stock market as a whole

Vejleder: Thomas Mikosch
Censor:    Mette Havning