Title: Platform market power in online labor markets: Evidence from a fee change
Room: E22
Time: 12.30
Abstract: We study how fees charged by an online labor platform affect the behavior of workers and employers, as well as the ecosystem as a whole. To achieve this, we use proprietary microdata from a large worldwide online labor platform for freelancing projects from 2014 to 2016. We exploit an unexpected change in the fees charged to new workers as a natural experiment to study the impact on individual worker behavior. Our findings suggest that workers who incur higher platform fees (treated workers) attempt to pass on the fee increase to employers by submitting higher bids for projects. This is especially the case for new, less established, and female workers. However, treated workers do not achieve higher wages and are less likely to win contracts, resulting in lower gross revenues. The higher fees lead to a further decrease in workers’ net revenues. The quality of work, as measured by employer feedback, and project duration are unaffected. We also explore the effect of higher platform fees on labor demand, labor supply, and equilibrium wages by instrumental variable regressions, confirming that a higher share of treated workers only affects labor supply, but not labor demand or equilibrium wages. We conclude that the fee change substantially increases profits by raising the share of platform fees on total turnover, without having significant side-effects on volume, quality and prices.