All That Glitters Is Not Gold: The Impact of Certification Test Costs in Online Labor Markets

Third-party skill certification is widely employed to address the ubiquitous information asymmetry between workers and employers. Offering certification tests at zero cost could be a natural strategy to encourage workers to certify their skills; however, neither theoretical nor empirical studies have examined how employers react to such a policy. This research fills that gap. First, we develop a stylized model to theorize the effects of free certification tests on employers’ recruitment decisions and the transaction amounts (i.e., contract prices) with workers at both the job and platform levels. Second, we empirically test these hypotheses using a natural experiment in which one of the largest online labor markets unexpectedly eliminated certification test fees. The policy change caused a surge in certification attempts and an increase in the number of certified workers, enabling employers to make faster hiring decisions. However, contrary to the platform’s expectations, the zero-cost certification policy diluted the signaling value of the certifications. Specifically, employers became less likely to prefer certified workers in terms of both hiring probability and contract price. Crucially, this change appeared to be economically justified because workers certified under the new policy were less likely to deliver high-quality work. Moreover, the decline in signaling value disproportionately affected inexperienced workers, who rely heavily on certifications to compensate for their lack of employer feedback. Finally, we demonstrate that by improving the accessibility of certification test information, platforms can better preserve the value of certifications while minimizing dilution effects. Our findings strongly support the study’s theoretical propositions and provide important insights for online labor market platforms and practitioners.

History: Hsing Kenneth Cheng, Senior Editor; Jason Chan, Associate Editor.

Funding: M. Lin acknowledges the generous financial support of the Ewing Marion Kauffman Foundation [G-201707-2578 and G-201812-5687]. An early version of this paper, titled “Information Unraveling in Online Labor Markets: Evidence from Natural Experiments,” was funded by the 2012 NET Institute Summer Research Grant [no. 12-02, awarded to M. Lin and P. Goes].

Supplemental Material: The online appendix is available at https://doi.org/10.1287/isre.2023.0303.