We estimate the effects of exchange rate on U.S. employment, exploiting differences in industrial composition across major cities. We find that a 1% depreciation of export‐weighted real exchange rate has a positive 0.98% direct effect on manufacturing employment. Its indirect effect on local nonmanufacturing employment rises with the size of the local manufacturing sector, consistent with the hypothesis that there exists a local spillover from the tradable to the nontradable sector. In cities with heavy concentration of manufacturing employment, the indirect effect is statistically significant and about 60% as large as the direct effect measured by the number of jobs. (JEL F3, F1, J2)