eThesis - Kauppakorkeakoulun sähköiset gradut

This thesis is concerned with exotic analysis methods for crude oil price formation process. It introduces new factors into academic discourse on the topic; mainly the political stability - price and political stability - inventories relationships. The potential for building models with forecasting power was also discussed. The statistical explanation power of the models were too weak, in order to implement a full-scale forecasting model. The data were analysed by OLS-methods with dummy variables, as well as with Dynamic Panel Data models, such Arellano-Bover/Blundell-Bond. The mathematical elegance, feasibility, and limitations to the models were discussed in detail. Firstly, the results indicate that the positioning of the sovereigns in the crude oil market signalled statistically significant information, for the time period 1990-2011. As the countries increased their crude oil inventories, the prices of the crude oils decreased. This held true for WTI, Brent and Dubai, but not for Urals. As Urals is extracted in the non-OECD country of Russia, the data is redundant in explaining the Urals price changes. Secondly, statistically significant evidence was presented regarding the open interest positions of the NYMEX traders in the WTI futures market, for the time period 1995-2011: the Commercials' did position worse than the other market participants. Further on, the Non-reportables positioned most often correctly in the market. Explanations for the phenomena were analysed - most notable were the balancing flows between the Commercials' and the Non-commercials' positions. Thirdly, statistically significant evidence was presented regarding the role of political risk as a price-setting force in the crude oil price formation process, for the time period 1992-2011: both, the IFO WES - Present and 6 Months indices, explained the increase in crude oil price when political stability deteriorated, and vice versa. The results applied to WTI, Brent and Dubai, but not to Urals. For the Urals, the curios symbiosis of politics and business in Russia was analysed as an explanation factor. Lastly, evidence was presented regarding the role of risk ratings as a defining factor for the amount of crude oil production in producing countries, for the period 2000-2011. As the countries' EIU Risk Ratings increased, the production levels decreased, and vice versa. The insignificance of the explanation power was discussed, and ideas for model development and the nature of the data were discussed: the model was robust, but the data included excessive noise for the purpose of a proxy in statistical inference. Also, the data illustrated a correlation-causation bias.