The Paradox of Search Frictions in Online Labour Markets

Authors
Ashley Pople

The labour market is increasingly “wired” with the proliferation of more accessible technologies, transforming both the nature of work and the ways in which firms find workers. However, low marginal application costs in these online markets have resulted in high volumes of ill-targeted job applications or “congestion” and poor match outcomes as a result. In light of this problem, this project addresses the degree to which design features of online markets help or hinder firms find the right workers. Despite the drive towards frictionless markets, the optimal level of friction in application decisions is unlikely to be zero. In this project, Pople will investigate whether introducing small application costs that vary in size and content attracts workers with better “job fit” and improves productivity through experiments with workers in South Asia on Upwork, the largest freelancing platform in the world. The study will also consider the impact small application costs could have for more marginalised jobseekers who otherwise struggle to signal their fit through traditional proxies (e.g. education certificates, assessments or brand names from employers), including women and entry-level workers.

The project will create fictional employer accounts and post jobs for an identical data-entry task to be completed over several days. Applicants are stratified by gender and work experience and 2000 are randomly hired to perform the task. Their job satisfaction and productivity are then measured and compared between four treatments, across which the job postings and respective successful applicants are randomised. These are: (1) a control group; (2) a minor attention cost group, in which jobseekers are instructed to include “interested” at the top of their cover letter in the job description, signalling an ability to pay attention; (3) a larger attention cost group, in which jobseekers are instructed to open an attachment in the job posting and answer a question about a detail included in the attachment, signalling an ability to pay attention and a willingness to exert effort; and (4) a screening questions group, in which jobseekers self-report their job fit, rather than showcasing it through their behaviour.

Online job matching and freelancer platforms have recently proliferated worldwide, with support from policymakers. This study is an opportunity to explore how best to design these markets and the effects of these design features on productivity and retention for firms. The research agenda was co-created in partnership with a job platform in Bangladesh and based on qualitative fieldwork, ensuring that the research is policy-relevant both within and beyond the country. This partnership offers an opportunity to scale any insights from the research. Further, the job platform has existing relationships with some of the largest employers and training providers in Bangladesh, providing a far-reaching network that will enable dissemination of results to maximise policy impact.

Authors

Ashley Pople

University of Oxford