The importance of model risk has been noted and investigated by several different researchers for the last decade, especially its application in the financial market. This thesis investigates theoretical spot prices from the nordic electricity market from 2006 to 2018 and attempts to predict the day ahead prices by the use of historical data and the GARCH-model with different assumed distributions for the white noise. Further on the different Value at Risk from the models are calculated and compared with the help of a relative measure of model risk.