Preference elicitation in generalized data envelopment analysis : in search of a new energy balance in Japan

The recent dramatic change in energy supply in Japan has prompted a search for a new energy-environment-economic efficiency policy, in which a compromise has to be found between a sufficient supply of energy resources, the development of low carbon emission technology, and a continuation of economic growth. The prefectures in Japan - 46 in total (excluding Tokyo) - are regarded as the institutional agents or decision-making units (DMUs) which are responsible for the design of a new sustainable energy balance in these regions. The main challenge is now to design an efficient energy-environment-economic system. The present paper aims to develop a balanced decision-support tool for achieving an efficient energy supply in all Japanese prefectures. To that end, a new variant of Data Envelopment Analysis (DEA) is presented, which is characterized by two integrated features: (i) the use of a general Euclidean Distance Method (EDM) to achieve the most appropriate movement towards the efficiency frontier surface (in contrast to the standard radial movement, leading to a uniform proportional input reduction – or uniform proportional output increase); (ii) the incorporation of preference-based (PB) adjustments in efficiency strategies regarding the input reduction allocation - or the output increase allocation - of DMUs in order to balance rigorous efficiency decisions with political priorities at the regional level. This paper illustrates this new methodology by means of an application to prefectural energy efficiency strategies in Japan.