Pricing swing options in the electricity market

Abstract: The thesis deals with how to price swing options in the electricity market by using a least squares Monte Carlo method. This is a simulation method which uses a backwards moving algorithm where the optimal decision is calculated at every time step. Regression is used for the optimal decision and in this thesis both a polynomial regression and a cubic smoothing spline are used. They are both shown to be rather good estimators for the regression. Two variation of contracts are priced. For the rst only one exercise right that can be used when exercising and for the second one several exercise rights can be used when exercising. Volume restrictions are also used. The algorithm implemented in this thesis give similar results to the ones of previous authors and when we can not compare with other authors it give us results fairly close to our expectations. The thesis also examines the optimal exercise strategy for a swing option and the boundaries for when to use an exercise right are calculated.