Technological Change, Occupational Structure, and Wage Inequality in India: The Role of Trade Unions and Labour Regulations

This paper examines how trade union density and employment protection laws mediate the distributional effects of exposure to artificial intelligence (AI) on wages in India's urban labour market. Using nationally representative data from the Periodic Labour Force Survey (PLFS) 2024-25, we construct an occupation-level Artificial Intelligence Occupation Exposure (AIOE) index and employ quantile regression alongside the Recentered Influence Function (RIF) Oaxaca decomposition to analyse wage inequality across the distribution. The results reveal that the impact of AI exposure is highly uneven and strongly conditioned by institutional settings. While workers in high-exposure occupations experience increasing wage premiums toward the upper quantiles, these gains are neither uniform nor automatic. Trade unions and employment protection laws play a critical role in shaping these outcomes. In particular, higher trade union density is associated with wage protection at the lower tail of the distribution, suggesting a compression effect that benefits low-wage workers. However, this effect is non-linear: unions appear to moderate or redistribute gains at the median, while their influence becomes more limited or heterogeneous at the upper end, where firm-level bargaining and skill complementarities dominate. More regulated (inflexible) labour market regimes provide measurable wage support for low-wage workers but are associated with weaker wage growth in the middle and upper segments, indicating a trade-off between protection and flexibility. Conversely, flexible regimes tend to amplify inequality by allowing stronger returns to AI exposure at the top, while offering limited protection at the bottom. Interaction results confirm that the wage effects of AI exposure are contingent on the joint configuration of trade unions and regulatory regimes. Our wage decomposition model further shows that institutional factors, particularly access to formal employment and social security, are central to explaining gender wage disparities, especially at the lower and middle quantiles.

Overall, the findings underscore that technological change is not inherently inequality-generating. Rather, its distributional consequences depend critically on the strength and design of labour market institutions. Strengthening collective bargaining mechanisms and extending employment protections are essential for ensuring inclusive outcomes in the era of AI-driven transformation. Our results are robust to alternative specifications.