We study this phenomenon using data from a novel job search platform called “Job Talash” in Pakistan. We recruited platform users through a representative survey of over 50,000 households in Lahore, Pakistan. At the time of sign-up, these jobseekers varied widely in their employment and job search status, ranging from employed and searching to non-employed and non-searching. This sample breadth is unusual in experimental job search studies (Poverty Action Lab 2022). Using the platform requires only basic literacy, a simple phone, and almost no airtime, generating very few technological and financial barriers to search.
Through the platform, we prompt employers every month to list any new vacancies. We match jobseekers on the platform to these new vacancies based on their education, work experience, and occupational preferences. When users match to a new job, we send them a text message inviting them to apply and then forward their applications to employers. This system generates rich data on both supply and demand sides of the labour market: job attributes, application and interview decisions, and how jobseekers update their CVs over time. This yields a very precise description of the job search process up to the interview stage.
We randomly assign some jobseekers to get a phone call after their text message. The phone call informs them about the jobs they’ve matched to and invites them to initiate an application, without providing any additional information or encouragement. Meanwhile, control group users must call the platform or ask the platform to call them to initiate job applications. Thus, the treatment only changes how jobseekers initiate applications on the platform, moving them from an active role to a passive role.