Wage Signaling in Online Freelancing Markets

How to cite

Wu, Ruoxuan; Ahmed, Hamna; Tirmazee, Zunia Saif. 2025. "Wage Signaling in Online Freelancing Markets." AEA RCT Registry. AEARCTR-0015840. https://doi.org/10.1257/rct.15840-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Ruoxuan Wu* , Hamna Ahmed , Zunia Saif Tirmazee

* Primary investigator

Registry ID
AEARCTR-0015840
DOI
https://doi.org/10.1257/rct.15840-2.0
First registered
Last updated
Status
In development
Trial dates
2025-04-21 – 2026-09-30
Countries
Private
JEL codes
J23, J24, J31, L15
Registry record
http://www.socialscienceregistry.org/trials/15840

Abstract

The rise of online freelancing platforms offers a promising avenue to reduce youth unemployment in low- and lower-middle-income countries by addressing typical constraints such as limited job opportunities, high search costs, and information frictions. Yet, unemployment remains high in these markets, particularly among inexperienced workers. While economic theory suggests that inexperienced workers can bid down wages to clear the market, wage undercutting is rare. This project explores the hypothesis that freelancers may choose not to engage in undercutting because proposed wages act as signals of unobserved quality. To understand wage-setting in online labor markets, we combine two RCTs to answer: 1) how signaling concerns and uncertainty about own skills affect wage-setting decisions of inexperienced online freelancers; and 2) how employers’ demand and beliefs about worker quality respond to wage undercutting.

Experimental design

Worker experiment: We measure the performance of inexperienced workers and refer them to a job opportunity. We randomly vary workers' signaling concerns and information about own ability in the job referral message.

Employer experiment: We take an audit study approach to measure consumer demand. We send applications to real online jobs and randomly vary the experience level of the worker profile and the proposed wages.

Design details

Randomization method
Randomization done in office by a computer.
Randomization unit
Worker experiment: worker level. Employer experiment: application level.
Clusters
n/a
Observations
Worker experiment: 400-500 individuals Employer experiment: 2000-3000 applications to 500-750 jobs
Arms
Worker experiment: evenly split among 4 treatment arms. Employer experiment: evenly split among 4 treatment arms.
Primary outcomes
Worker-side: proposed wages, actual performance, beliefs about own performance. Employer-side: response to job applications (interview request, job offers, etc.)
Secondary outcomes
Worker-side: decisions to apply for jobs/accept job offers, other application characteristics.

Cited by

No citation data available for this registration yet.

Keywords