Download
Abstract
We study how coworker networks shape industry mobility through two channels: the accuracy of workers’ wage beliefs about other sectors and their perceived probabilities of receiving job offers. Current coworkers improve wage belief accuracy: having all versus no current coworkers with prior experience in a given industry reduces median-wage prediction errors by 29 percent of their mean. Both current and past coworkers raise perceived job-offer probabilities and stated application intentions. Using novel survey data from a representative sample of about 3,000 U.S. wage and salary workers, we find that having all past coworkers employed in a given industry increases perceived offer probabilities by 11 percentage points (a 29 percent increase relative to the mean) and raises application intentions by 29 percentage points, while the same shift for current coworkers raises both by 17 percentage points. A conjoint experiment confirms that active coworker connections raise perceived offer rates by 11.4 percentage points, more than any other job attribute in our design. We embed these findings in a structural sector-choice model calibrated to CPS transition flows and survey moments to quantify the aggregate implications of coworker influence for labor market mobility.