Exporting pollution?

Economic activity causes to varying degree pollution to air, soil and water. The pollution that is caused by the production of a single unit of a certain commodity can be said to be embodied in the commodity. This approach allows us to investigate environmental issues from a consumption-centred perspective, and this is especially important in the case where the embodied pollution in a certain commodity is not restricted to the country where the commodity is consumed.

In this thesis, I present some possible approaches to measure the emissions embodied in the exports and imports of different commodities. The indicator which I focus on, the so-called Pollution Terms of Trade is presented in the second chapter, and its main features is discussed and I also propose several possible applications. The indicator is based on trade data and emission intensities in the production of different commodities, and I pay particular attention to the importance of using country-specific intensities in order to capture the total embodied emissions in trade.

In the third chapter I calculate the indicator for the case of Norway, and find significant differences in the balance of emissions in trade for different pollutants. The importance of the oil industry, as well as ocean transport, for Norwegian exports greatly affects the results, especially in the case of pollutants like NOX and NMVOC. The technology effect is also significant in some cases, as expected.

In the fourth chapter I discuss the impacts of trade on the environment, and present a theoretical model to illustrate some of these issues. I also use data from an earlier general equilibrium study to assess the effects of a certain shift in trade policy on the Pollution Terms of Trade. I find that as the net import of agricultural goods increases, the balance of embodied emissions in trade improves to a great deal when it comes to pollutants like CH4 and NH3. The effects are both positive and negative in terms of the PTTI, depending on type

Acknowledgement : I would like to thank my main tutor, Taran Fæhn for invaluable guidance throughout the process. In addition I would like to thank Jon Strand, Annegrete Bruvoll and Birger Strøm for useful comments and information. Several other staff at Statistics Norway have also been helpful in the process of collecting data, and the institution has provided me with a excellent work environment.

The study has been financed by Statistics Norway and The Norwegian Research Council (The RAMBU program).