How much do employees' wages directly reflect their employer's labor demand, rather than competition from other employers in the labor market? We test the wage incidence of product demand shocks by studying a quasi-experiment that idiosyncratically shocked individual firms' export demand without systematically affecting similar firms' product or labor demand. Our shocks measure how much Portuguese exporters' sales were impacted by where—but not what—they had been selling before the recession of 2008. These shocks predict changes in output, payroll, and hiring at affected firms, but not at rival employers in the same labor market segment. An idiosyncratic shock that changes output by 10 percent in the medium-run causes wages of pre-2008 employees to change proportionally by 1.5 percent, relative to trend. Consistent with a simple framework, we find that these pass-through effects are larger in industries with lower employee turnover rates and in firms with higher pay premiums. These findings offer evidence that heterogeneous firm dynamics can plausibly generate substantial cross-sectional wage dispersion, but only in less-fluid labor markets.