Estimating the efficiency of American petroleum refineries under varying assumptions of the disposability of bad outputs

Purpose - The paper aims to describe and compare multiple methods for estimating the technical efficiency of 113 US oil refineries in operation in 2006 and 2007, considering undesirable output in a production process. Design/methodology/approach - A technology that satisfies weak disposability between desirable and undesirable outputs is constructed by allowing different abatement factors across all refineries. Several measures based on data envelopment analysis approaches are implemented and compared to study the impact of disposability assumptions and to investigate the effects of using non-uniform abatement factors. A hyperbolic efficiency measure is used to analyze the potential output loss of each refinery due to environmental regulations. Findings - The results indicate that domestic refineries can improve efficiencies regardless of the disposability assumptions and that environmental regulations reduce the amount of potentially desirable outputs produced by some facilities. However, refineries in the western USA appear to be the most affected by regulations. In general, efficient refineries are less likely to be affected. Research limitations/implications - Undesirable outputs are limited to toxic release. Undesirable outputs generated from refining crude oil, such as greenhouse gases, can be used when data are available. The desirable outputs in this paper do not include premium products, such as lubricants, which could raise the efficiency estimates of complex refineries. Originality/value - To the authors' knowledge, this paper is the first implementation of the weakly disposable technology constructed by different uniform abatement factors. Further, the paper investigates the effects of various disposability assumptions on efficiency estimation. The result clearly identifies refineries that use their resources efficiently. The paper suggests that the data may be used to augment managerial decision-making regarding benchmarking and best practices.