An important and illegal source of monopsony power are non-compete clauses in workers’ contracts that avoid employers bidding up wages. This paper presents a framework in which the source of monopsony power also originates from workers’ contracts without violating anti-trust regulations: contracts offered by labor market intermediaries such as temporary help agencies. Building on the static monopsony model of Card et al. (2018), we show how intermediation results in wage markdowns and why some workers consent with being paid below their productivity. We then show evidence in support of our model using administrative matched employer-employee data for workers who are being payrolled by their employer.

We examine the impact of job search assistance for the unemployed on their search effort on an Online Job Platform (OJP). We find that individuals who search more intensively without assistance are also more likely to take-up assistance when offered. Using a staggered DiD design that allows for selective take-up, we find that those more likely to take-up assistance are also more responsive to it. This implies that barriers to job search assistance effectively screen out those that need it most. We also find that the impact of assistance is on search effort is short-lived, and that there is strong negative duration dependence in job search effort.

This paper shows that technological progress also has substantial impact on the nature of competition in labor markets, and that accounting for this impact results in a better understanding of technology-induced changes in workers' relative wages. The paper first presents a simple monopsony model in which firms have wage-setting power because workers are in excess supply, and in which technology not only increases the marginal product of labor but also decreases firms' wage-setting power. As a result, not only average real wages grow but lower-tail wage inequality also decreases, driven by stronger relative wage growth in the lowest-paying firms and a relocation of workers from the lowest-paying to higher-paying firms. These predictions are then tested using unique data between 1846 and 1896 during Belgium’s First and Second Industrial Revolutions, when trade unions and other institutions protecting low-wage workers did not yet exist.