Browsing by Author "Lehto, Tomi"

In this thesis the influential oil market model proposed by Kilian and Murphy (2014) is studied with statistical identification. The identification method used provides point identification in the set defined by restrictions used in the earlier literature. This method allows us to obtain point estimates for the price elasticity of oil production which has been the source of differences in the results of the earlier literature. Additionally, the importance of oil market shocks in explaining the changes in oil prices is studied with the identified models. The earlier literature suggests that oil demand shocks are more important drivers of oil prices than oil supply shocks. Studies allowing for higher levels of price elasticity of oil production find oil production shocks to be relatively more important than the studies using low upper bounds for this elasticity. According to microeconomic studies, higher levels of price elasticity of oil production could be explained by increases in shale oil production. Identification of the oil market SVAR model is achieved by extracting the model which maximizes the independence of the structural shocks. This optimization task can be challenging. Hence the performance of the estimation can be improved with prior knowledge about the structural shocks by inclusion of restrictions into the optimization problem. In this thesis the distance covariance measure is used to measure the independence of the structural shocks. The distance covariance measure is minimized to obtain the model with the most independent structural shocks in the set defined by the restrictions. The constrained optimization problem is solved with external penalty optimization method. Two distinct minima are obtained from the optimization task which suggest different dynamics for the oil market shocks. The first optima suggest that the price elasticity of oil production is close to zero. The other optima suggest higher levels for this elasticity. Both optima suggest that the elasticity of oil production differs in response to different shocks. With the first optima supply shocks are found to be more important drivers of oil prices than found in the earlier literature. With the second optima the importance of production shocks is found to be more similar to the results obtained in the earlier literature. Additionally, speculative shocks are found to be important drivers of the oil prices with this optimum. The method which uses external penalty optimization method is compared to the method based on sampling used earlier in the literature. It is found that the method based on sampling finds less independent structural shocks than the method based on external penalty optimization method.