A STATISTICAL MODEL FOR ESTIMATING NATURAL RESOURCE DAMAGES FROM OIL SPILLS

摘要

This paper presents a statistical model using multiple-regression analysis that explains variations in past natural resource damage (NRD) settlements (excluding assessment costs) for oil spills in the United States based on a variety of factors, such as the amount of oil spilled, the type of oil spilled, and the geographic location of the spill. The results of this statistical model indicate that, other things being equal, NRD settlements are higher for larger oil spills, spills in California, spills that occur in the winter months (December, January, and February), and when endangered species are injured. Alternatively, other things being equal, NRD settlements are lower for spills of light crude oil than for other types of petroleum products (mainly heavy crude oil), and spills with unvalued compensatory restoration projects. In this paper, the authors use their statistical model to predict a point estimate and range for the NRD settlement for two hypothetical oil spills. This demonstrates that their statistical model could be used to predict the NRD settlement for future oil spills within a few weeks of occurring, which could significantly shorten the time required to reach a settlement.