Disputas: Linda Vos
M. Sc. Linda Vos ved Matematisk institutt vil forsvare sin avhandling for graden ph.d.:
Stochastic volatility and multi-dimensional modeling in the European energy market
Tid og sted for prøveforelesning
Lecturer Luitgaard Veraart, Department of Mathematics, London School of Economics and Political Science
- Associate Professor Sjur Westgaard, Institutt for industriell økonomi og teknologiledelse, NTNU
- Professor Giulia Di Nunno, Matematisk institutt, Universitet i Oslo
Leder av disputas
Professor Bernt Øksendal, Matematisk institutt, Universitet i Oslo
- Professor Fred Espen Benth, Matematisk institutt, Universitet i Oslo
- Professor Koekebakker Steen, Institutt for økonomi, Universitetet i Agder
In energy prices there is evidence for stochastic volatility. Stochastic volatility has effect on the price of path-dependent options and therefore has to be modeled properly. We introduced a multi-dimensional non-Gaussian stochastic volatility model with leverage which can be used in energy pricing. It captures special features of energy prices like price spikes, mean-reversion, stochastic volatility and inverse leverage. Moreover it allows modeling dependencies between different commodities.
The derived forward price dynamics based on this multi-variate spot price model, provides a very flexible structure. It includes cotango, backwardation and hump shape forward curves.
Alternatively energy prices could be modeled by a 2-factor model consisting of a non-Gaussian stable CARMA process and a non-stationary trend models by a Levy process. Also this model is able to capture special features like price spikes, mean reversion and the low frequency dynamics in the market. An robust L1-filter is introduced to filter out the states of the CARMA process. When applying to German electricity EEX exchange data an overall negative risk-premium is found. However close to delivery a positive risk-premium is observed.
For mer informasjon
Kontakt Matematisk institutt.