Scale, Technique and Composition Effects in Manufacturing SO2 Emissions

Combining two data sources on emissions with value-added and employment data, this paper constructs six data bases on sulfur dioxide (SO2) intensities that vary across countries, sectors and years. This allows us to perform a growth decomposition exercise where the change in world manufacturing emissions is decomposed into scale, composition and technique effects. The sample covers the period 1990-2000, and includes 62 countries that account for 76% of world-wide emissions. While manufacturing activity has increased by a rough 10% (scale effect), we estimate that emissions have fallen by about 10%, thanks to the adoption of cleaner production techniques (the technique effect) and a small shift towards cleaner industries (between-sector effect). As output and productivity gains have been biased towards large emerging countries like China and India, which are both clean in terms of emissions per unit labor and dirty in terms of emissions per dollar, the sign and magnitude of the between-country effect depends on the choice regarding the scaling factor ( - 2% for employment, + 25% for value-added, with a corresponding adjustment of the technique effect). The paper also shows that these estimates are robust to changes in aggregation across entities (regions or countries) and across industries, and that composition changes are correlated with changes in prices and trade intensities.