WAGE INEQUALITY AND SKILL DEVELOPMENT IN INDIA

Abstract

India's labour market has gone through considerable restructuring over the past two decades. Economic liberalisation, rapid technological change, and an expanding service sector have together produced a labour market in which skilled workers capture a disproportionately large share of wage gains, while low-wage workers, who make up the vast majority of the workforce, continue to face stagnant real earnings and precarious employment. Against this backdrop, successive central government administrations have invested heavily in skill development programmes, from the Industrial Training Institutes (ITIs) to the Pradhan Mantri Kaushal Vikas Yojana (PMKVY) and the Deen Dayal Upadhyaya Grameen Kaushalya Yojana (DDU-GKY), disbursing over ₹6,000 crore annually in an effort to close the skill gap and improve labour market outcomes for disadvantaged workers. What remains thin, even so, is robust, distribution-sensitive evidence on whether these interventions actually reach the workers who need them most. This dissertation looks at the relationship between skill development and wages across the full wage distribution of Indian workers, using data from the Periodic Labour Force Survey (PLFS) 2023 to 24, the most recent nationally representative household survey conducted by the Ministry of Statistics and Programme Implementation, Government of India. The analytic sample comprises 46,294 wage workers aged 15 to 65 drawn from both rural and urban areas across all major states. The core estimation strategy is quantile regression (Koenker & Bassett, 1978), which allows the conditional wage effects of education, vocational training, and other worker characteristics to be estimated at five quantile points, the 10th, 25th, 50th, 75th, and 90th percentiles of the log-wage distribution, directly addressing the main limitation of prior Indian studies that rely on OLS and estimate only mean effects. The econometric model is grounded in an augmented Mincerian earnings function. The dependent variable is the natural log of the monthly wage. Key regressors include years of formal education, a binary indicator for formal vocational training (skill dummy, equal to 1 if the worker received certified vocational training), potential labour market experience constructed as age minus years of education minus 6, experience squared, and a set of demographic controls for gender (female dummy), rural/urban location (urban dummy), and social group membership (SC, ST, and OBC dummies), with 'Others' as the reference category. Employment type (self-employed, regular wage, casual labour) is also incorporated in the OLS specifications. Robust standard errors are used throughout. Six main findings emerge. First, returns to formal education rise monotonically across quantiles, from 6.9 per cent per year at the 10th percentile to 11.5 per cent at the 90th percentile, confirming that education is a force for wage divergence at current rates of return. Second, vocational training yields an inverted-U pattern: returns peak at the median (18.2 per cent), are moderate at the lower tail (11.4 per cent at Q10), and decline at the upper tail (10.5 per cent at Q90), making training more egalitarian in its distributional footprint than formal education. Third, the female wage penalty is largest at the bottom of the distribution (−67.9 log points at Q10) and narrows substantially but remains large at the top (−19.4 log points at Q90), consistent with the sticky floor hypothesis. Fourth, SC and OBC wage penalties widen at higher quantile points, indicating that caste-based barriers to high-skill, high-wage employment are the primary mechanism of disadvantage rather than uniform wage suppression. Fifth, the urban wage premium is largest at Q10 (23.2 per cent) and declines modestly to Q90 (17.1 per cent), suggesting urban labour markets provide proportionately greater gains to low-wage workers. Sixth, experience returns are stable across quantiles at approximately 4 per cent per year and do not contribute to distributional divergence. These findings carry direct policy implications. Targeting PMKVY and DDU-GKY enrolment toward low-wage workers, particularly women, SC/ST workers, and rural youth, would concentrate public resources where the relative returns to training are meaningful. Reforming curriculum and certification to align with industry demand, expanding training infrastructure in rural areas, designing gender-disaggregated programmes, and strengthening anti discrimination enforcement in higher-wage employment are complementary interventions that training policy alone cannot substitute.