Input–output analyses of the pollution content of intra- and inter-national trade flows

Abstract

This paper considers the application of input–output accounting methods to consider the pollution implications of different production and consumption activities, with specific focus on pollution embodied in intra- and inter-national trade flows. It considers the illustrative case studies of production and consumption measures of emissions and pollution embodied in interregional trade flows between two regions of the UK and between five Midwest regions/states within the United States. The analysis raises questions in terms of policy reliance on the extremes of conventional production and consumption accounting measures and considers a range of alternative measures that may be calculated using input–output methods to provide different informational content. The paper focuses on different types of air pollutant of current policy concern in each the UK and the US Midwest cases and demonstrates how use of the environmental input–output framework allows the analysis of the nature and significance of interregional pollution spillovers. The results raise questions in terms of the extent to which authorities at the regional level can limit local emissions where they are limited in the way some emissions can be controlled, particularly with respect to changes in demand elsewhere within the national economy. This implies a need for policy coordination between national and regional level authorities to meet emissions reductions targets. Moreover, the existence of pollution trade balances between regions also raises issues regarding net losses/gains in terms of pollutants as a result of interregional trade. In conducting analyses for different types of air pollutant (here carbon dioxide, CO2, as a global warming gas, a greenhouse gas (GHG), in the UK case; and ammonia, NH3, as a pollutant of more local concern in the US case) the paper also considers how pollution embodied in international trade flows may be accounted for and attributed.