Efficient waste and pollution abatements for regions in Japan

Abstract

This paper computes efficient industrial waste and air pollutants abatements for 47 regions in Japan for the period 1992–2002. The variable-returns-to-scale (VRS) data envelopment analysis (DEA) with a single output (real GDP) and seven inputs (labor, real public capital stock, real private capital stock, industrial waste, sulfur oxide, nitrogen oxide, and soot and dust) is used to compute target wastes of each region for each year. The efficient abatement ratios of each region in each year are obtained by comparing the actual to the target amount of a pollutant. Our major findings are: (1) Most regions in Japan have significant room to reduce their pollution since there is a wide gap between efficient and inefficient regions; (2) For each air pollutant, approximately 25–33% of Japan's prefectures can reduce their output by more than 50% without harming regional GDP, and approximately one-third of prefectures can reduce industrial waste more than 30%; (3) Hokkaido is the least efficient region for all years studied and for all waste and pollutants, and target abatement ratios there drastically worsened in the last two sample years; (4) Tokyo, Saitama, Yamanashi, Shiga, Nara, and Tottori are efficient with respect to each type of industrial waste and pollution throughout the study period; (5) many regions in the bottom quartile with respect to real per capita income have significant room to reduce their waste and pollution output; and (6) many regions where energy-intensive industries dominate produce excessive amounts of waste and air pollution compared to other regions.

Keywords:

• data envelopment analysis
• efficiency frontier
• abatement targets
• industrial waste
• air pollution

Acknowledgements

The authors thank seminar participants at the All China Economics International Conference, East Asian Symposium on Environmental and Natural Resource Economics, and Asia-Pacific Productivity Conference. The authors are also indebted to Max Spoor and Chris Miller for their helpful comments. Partial research grant from Taiwan's National Science Council (96-2415-H-009-002-MY2) is gratefully acknowledged.