| dc.description.abstract | Resume
"How do oil prices affect oilrig activity? An empirical investigation" by Guro Børnes Ringlund. Supervisors: Knut Einar Rosendahl and Terje Skjerpen.
In this thesis, I analyse the relationship between oilrig activity and oil price changes for several oil-producing regions in the world. Rig activity is a preparation for future production of oil, through exploration for new fields or development of existing fields, and is thus an indicator for the future level of oil production. Few studies have discussed this relationship before. I want to find out whether oil price changes have any effects on oilrig activity in the short run, and what are the effects in the long run. Some of the regions include OPEC-countries, and as OPEC rather are price-makers than price-takers, these regions have also been estimated with the OPEC-countries removed, to see if this gives any different effects.
I use oilrig data from the Baker Hughes Rig Counts, whereas the price data have been collected from Petroleum Intelligence Weekly. The estimation period is 1992:1-2002:7 for the US region, and 1995:1-2002:7 for the rest of the regions. Since the data are time series data, I test the variables for stationarity, and find that they are integrated of order 1 (I(1)). Thus, to obtain valid inference, the variables need to be cointegrated, and I test this by way of the ECM-method (Equilibrium Correction Model). I find that the variables are cointegrated for most of the regions.
The thesis is organised as follows: Chapter 1 gives an introduction to the topic, and presents some facts about the oilrig market and the oil market in general. Chapter 2 presents the econometric theory of non-stationarity, unit roots and cointegration. I also present the concept of a stochastic trend, which turned out to be important for the cointegration properties of the variables. Chapter 3 gives a more thorough description of the data, and presents the empirical model used in the analysis. The point of departure is an ADL-model (Autoregressive Distributed Lag Model), which is reparameterised into an ECM-model (Equilibrium Correction Model), as this makes it easier to test for cointegration. I also introduce a stochastic trend in the empirical model. Chapter 4 reports the results of the estimations. First I discuss the short-term effects of oil price changes on the oilrig activity for each region, and then I go on to discussing the long-run price elasticities, and compare the size of these for the different regions. I find that, in the short run, only the US, Latin America and Non-OPEC Middle East regions show significant reactions to oil price changes, whereas in the long run, the US has a long -run price elasticity of 1.5, which e.g. is more than twice the size of the European long-run elasticity, at 0.6. A crude oil price increase from $25 to $30 (20 per cent increase) will thus in the long run lead to a 30 per cent increase in oilrig activity in the US, whereas the same price increase only will give a 12 per cent increase in European oilrig activity. Possible explanations for this may be the different degree of governmental involvement in the different regions (governmental regulation may delay the decision-making process), and varying flexibility of the rig markets (changing the number of rigs may be difficult in some regions, due to long-term contracts or lack of "spare" rigs). Also, for a couple of the regions (most notably Africa), it proves difficult to come up with a model specification that works. Finally, Chapter 5 concludes.
The model is estimated in STAMP 6.2. Unit root tests of the variables and evaluation of the stability of the dynamic models are performed in TSP 4.5. | nor |