The effect of spatial agglomeration (localisation and urbanisation economies) on employment growth is explored over a balanced panel of 23,374 Italian firms, using a flexible Bayesian model. Contrary to previous research, the agglomeration economies measures are calculated using direct measures of physical distances between pairs of firms, rather than with respect to pre-specified geographical units. We find that localisation effects are positive but decreasing with distance, while the variety effects are negative for distances within 10 kilometers and become positive for distances in a range of 10–30 kilometers. Our results suggest that the use of geographic units such as standard metropolitan units, LLS, administrative regions or provinces can be misleading.
- knowledge spillover,
Available at: http://works.bepress.com/claudio_lupi/11/