Do Online Certifications Improve Job Market Outcomes? Evidence from an it Skills Certification Platform in India

Abstract

We estimate the returns to online skills certification for engineering graduates in India using a regression discontinuity design. Individuals in our dataset take a computer programming exam, and are required to score above a threshold to receive a widely used software engineer IT-services certificate provided by an online job skills credentialing platform in India. We find that certified candidates have approximately double the odds of finding employment following the exam. Our results indicate that skill certification in this context is a strong one-time signal of quality. Certification cannot replace education, however, and fails to predict longer-term job market outcomes as demonstrated by the lack of estimated causal impact on current employment status or income level. Our results also demonstrate that an individual’s logical reasoning abilities and caste identity are significantly associated with long-term job market success. Together, these suggest the promise and limits of certification in an educational market where university training is highly variable.

Keywords: Regression Discontinuity, Certification, Skilling, Signalling, Computerised based assessments