The COVID-19 crisis has triggered unemployment worldwide. In developing countries, the effects are expected to be large (Mubarak and Howell 2020; Saleh and Cash 2020; Dingel and Neiman 2020); however, better data is needed to understand (1) who is most affected; (2) which job losses are expected to be temporary versus permanent, and (3) how to speed the recovery.
To address this, the research project builds on ongoing data from over 1,000 firms and 9,000 jobseekers in Lahore, Pakistan collected as part of a job search service developed by the research team, Job Talash. Both jobseeker and firm enrollment in the service were initially recruited from representative door-to-door samples across the metropolitan area, allowing us to characterize the distribution of effects across firms of varying industries and sizes, and for job seekers across education levels and by gender. The Job Talash system generates extensive high-quality survey and administrative data on both the supply and demand sides of the labor market. To date, the platform has generated administrative data for over 800,000 job –jobseeker pairs, one for each ad sent by the platform to a job seeker who meets the basic qualifications for that job. The project teams complements this with phone surveys of applicants and firms to measure outcomes on the supply and demand side, including off-platform search and hiring. They propose to build on this platform to achieve two key objectives.
Characterizing the distribution and duration of impacts: The researchers propose to build into this panel by adding survey waves to collect additional data from both firms and jobseekers on firm temporary and permanent closure, layoffs, furloughs, unpaid or reduced wages, and anticipated and actual resumption of work and search. With funding support for this project, existing survey teams who have already been trained in this project can be retained and quickly deployed to conduct these surveys over the phone. The project will randomize the order in which respondents are called over the coming months and merge it with previous and future data generated by the Job Talash platform on applications, interviews and employment on the same individuals and firms. This will allow us to characterize the trajectory of shock and recovery. The project will provide estimates of these statistics at different stages of the pandemic by subgroups, in particular by gender, wealth proxies, industry, and firm size.
Testing information on job opportunities to speed recovery: An ongoing randomized experiment increases the flow of information to job seekers on job opportunities via more frequent telephone outreach. The first stage results from this experiment are strong. The proportion of jobseekers whoever apply on the platform increases from 20 to 45%. Data collected to date on interviews also indicate a significant increase in interviews. At baseline, most firms and jobseekers searched primarily through social networks, which may often involve in-person meetings, or asking around at work sites, restricted during lockdown; in contrast, information through the platform is convenient to access even under social distancing. Experience searching over the platform pre COVID, and increased flow of information on job opportunities as hiring restarts, might help job seekers who have lost work, or their family members, find jobs sooner and recover from the economic shock faster.
Gender focus: The Job Talash platform was designed for experiments on female labor force participation (Subramanian 2020, Field and Vyborny 2020); thus, recruitment was designed to represent men and women, while women may be under-represented in other samples in employment studies (e.g. from job placement centers). In addition, our baseline data includes rich data relevant to gender, including unusually rich data on firms’ gender composition, facilities for female employees, and revealed willingness to recruit women, as well as the composition of earners in jobseekers’ households. This allows us to carefully characterize the distribution of the impacts by gender. The project will collect data on household-level job loss to help understand within-household effects, such as a male earner’s job loss increasing future job search by female household members to compensate for lost income. Capacity building: Researchers at the Center for Economic Research in Pakistan are actively involved in the intellectual development of this project. Duke-based researchers have worked to build capacity at CERP and in Pakistan, including delivering technical training for CERP staff and teaching on CERP Evidence-Based Policy Design workshops for policymakers. Duke and CERP researchers have co-authored project outputs (Sajjad et al. 2018, Delcuvellerie et al. 2019), and the project plans to co-author outputs from the proposed project. Policy outreach: An ongoing CERP-wide initiative engages with government agencies on all aspects of the policy response to the pandemic, building on CERP’s existing strong relationships with government partners; CERP is an official member of the working group of the Government of Punjab on COVID response. The project will engage with policymakers through these platforms throughout the project.