Over the last decade, the rapid diffusion of internet technologies in lower-income countries has changed how employers and workers interact in the labor market. While the initial boom in online recruitment focused on higher-wage occupations, there now exist online search and matching services for lower-wage occupations, from domestic work to entry-level retail jobs. Harnessing these data can provide real-time labor market insights that are both difficult to obtain through survey-based approaches and highly relevant for policy decision-making during the current COVID-19 crisis.
In this project, it outlines a plan to first assemble unique data on online vacancies in India and then use a range of empirical methods, including machine learning techniques, to understand changes in online recruitment behavior due to COVID-19 and the degree to which such changes can predict changes in the overall labor market. If online recruitment behavior is predictive of the overall labor market, the project team plans to develop an online interactive dashboard with results from our analyses. An additional benefit of collecting these data will be to understand to what degree gendered language in job postings can inform us about gender-based exclusion in the labor market and whether the current crisis will further exacerbate this exclusion. Such data would be particularly valuable in India, as it currently lacks any reliable infrastructure for collecting real-time labor market data and was already suffering from declining female labor force participation before the crisis.
Our data collection plan has two components. First, they will leverage an existing partnership with QuikrJobs, one of India’s largest online job search portals, to obtain their universe of available job postings from 2015 to the present. Second, they will collate data from other sources where vacancies may be posted (e.g. job search portals, large company websites, etc) by using Internet archives available using the Wayback Machine and build scraping algorithms to continue collecting these data in future months.3These two approaches will give us a fairly comprehensive dataset on online vacancies across regions, sectors, and occupations, and over time. For example, QuikrJobs has extensive reach in India, both sectorally and geographically, with millions of active users in over 1000 cities and 100+ occupations.