Arbnora Reci

Examing what drives the electricity price and forecasting eleectricity
price on a long term horizone

Abstract

Until the beginning of the 1990's the electricity price was determined
by the government in each Nordic country. To encourage a more open and
liquid market the Nordic countries decided to deregulate their electricity
market. The deregulation have changed the conditions on the Nordic
power market, the price has become more volatile and has grown
exponentially.
In this thesis we study the Swedish power market, where we examine
what drives the electricity price, traded on Nord Pool. The second purpose
of this thesis is to model and forecast the electricity price. Since
the forecast horizon is on a long term basis, twenty years, we model on
monthly basis. Forecasts on the electricity price are useful in multiple of
areas, such as in creating trading strategies, production planning and
predict
the demand of wind turbines. A linear regression model and a time
varying parameter regression models are used to examine the relationship
between the price and different explanatory variables. Since the price is
a market price, which means it's determined by supply and demand and
which varies along time we have a non linear relation. By adapting an
AR-model with bootstrapped external signals different forecasts of the
future
electricity price where made. The forecasts covers different scenarios
which are important when further analyzes are made based on that data.