*Trade Shocks** and Local Employment Multipliers: Evidence from France

In this paper, I develop a simple model of spatial equilibrium to investigate theoretically what determines the sign and magnitude of ``local multipliers' (defined as elasticity of employment in the non-tradable sector with respect to increase in employment in the tradable sector). I estimate the local multiplier with data for France and focus on the heterogeneity of local multipliers across local labor markets and across sectors. In order to cope with possible endogeneity issues, I use a shift-share instrument (Bartik,1991), already used in the literature, and build another instrument based on trade-shocks (in the spirit of Autor et al. (2012) and referred to as ``import-per-worker' index (IPW) below). I find an average elasticity of 0.36 (1.25 job-to-job effect), a result considerably higher than previous studies of European economies and similar to previous findings with the US data (Moretti and Thulin, 2012). The two instruments give broadly comparable results. I find that the local multiplier is larger for local labor markets with high initial unemployment rate. Similar results hold when using hours supplied instead of mere headcount, suggesting that 'tight' labor markets fail to adjust through the intensive margin. On the other hand, when using overall payroll, the difference between slack and tight labor markets cease to be significant. This suggests that increase in labor demand push up wages in tight labor markets while the adjustment occurs purely through quantities purely in slack labor markets. Hence the incidence of shifts in local labor demand depends on initial labor market conditions. Finally, I find evidence of higher local job multipliers associated with high-tech sector and show that when estimating the local earning multiplier the difference becomes insignificant, suggesting that more intensive pecuniary externality drives that sectoral heterogeneity and not some sort of technological externality.