Working Paper: CEPR ID: DP17674
Authors: Marc Ivaldi; Emile Quinet; Celia Ruiz Mejia
Abstract: The objective of this paper is to estimate the extent of agglomeration externalities taking into account the direct and indirect impacts of transport exposure on productivity. To do so, we take advantage of a rich data infrastructure that combines very fine georeferenced infra-municipality data on more than one million employees with detailed data on the public-transport and road networks of a typical European metropolitan area, namely the Toulouse Metropolitan Area (TMA). We recover the productivity effects of agglomeration and transport measures by the implementation and estimation of a wage determination model in two stages. The first stage assesses the importance of industrial concentration and employees’ characteristics against true productivity differences across zones on the average of local industrial wages. The second stage explains local productivity differences on our local factors of interest: agglomeration and transport. Finally, and to have a full representation of transport impacts, we investigate the size of the indirect effect of transport exposure on productivity by its impact on the distribution of metropolitan employment. We exploit the panel nature of our data and apply instrumentation techniques to cope with the endogeneity of agglomeration and transport measures. Our results suggest that both agglomeration and transport exposure measures have a substantial and significant effect on local productivity. Indeed, when density of employment doubles, productivity increases by 1.6%. Further, the effects of transport exposure measures differ for the two modes considered, private vehicle and public transport. In both cases, a higher exposure to transport supply implies higher levels of employment an productivity.
Keywords: agglomeration economies; accessibility; transport exposure; public transport network; road network; productivity; transport infrastructure; density; cities; commuting costs; urban economics; transport economics
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| local employment density (R23) | productivity (O49) |
| accessibility to employment via road network (R42) | productivity (O49) |
| accessibility to employment via public transport network (R41) | productivity (O49) |
| decrease in average optimal commuting times by road network (R41) | productivity (O49) |
| increase in travel times (R41) | productivity (O49) |