Productivity spillovers among linked sectors

This paper estimates the impact of inter-sectoral linkages on productivity at the sectoral level. An exhaustive Chinese panel data set for capital, infrastructure and a sectoral agglomeration index is linked with an economic distance matrix derived from inter-sectoral transactions. The latter matrix can replace the conventional geographic distance matrix from spatial econometrics. The impact through spillovers is mixed-the direct impact passing to related sectors and back to the initial sector itself, and the indirect impact arising from changes in all sectors. The results suggest that (1) economic growth in a sector is driven by spillovers among sectors that are linked through flows of goods and services; economic distance plays a more important role in stimulating productivity spillover than spatial distance; a shorter economic distance transmits a larger productivity spillover between sectors; (2) infrastructure spillover improves labor productivity in linked sectors; (3) agglomeration diseconomies can be partially reduced by infrastructure investment.

► Intersectoral econometric models are constructed and compared with non-intersectoral models and spatial econometric models. ► Economic distance is calculated based on intersectoral transaction and Leontief inverse matrix. ► Spatial data and economic data of 36 sectors, 31 provinces in 12 years are applied. ► Input-output linkages have both direct impact and indirect impact across Chinese sectors. ► Intersectoral spillovers have positive effects on productivity, although their magnitudes are negatively correlated with economic distance.