Can Matching on Worker Skills or Preferences Yield Better Firm-Worker Matches?

Developing country labor markets are often inefficient, with relatively long search durations, low-quality matches between firms’ needs and workseekers’ skills, and high turnover. Existing evidence shows that generating new information about workseekers’ skills increases these workseekers’ employment rates and earnings. We build off this work by matching applicants to entry-level vacancies using new and different types of information. We randomly vary, at the branch level, the criteria a hiring agent uses to shortlist candidates across hundreds of South African firm branches. We examine effects of shortlisting candidates using (1) worker preferences for job types or (2) communication and socio-emotional skills compared to (3) education and experience. We measure match quality, turnover, labor market access for disadvantaged groups, and branch-level hiring. We extend existing work by considering information about both skills and preferences and by examining on-the-job productivity, turnover, and labor demand.