AI-driven Technological Progress, Finite Change and Wages

When technological progress uniformly raises the productivity of all factors of production within a competitive industry, it typically increases the real returns to those factors-including the wage rate in that industry. This mechanism plausibly reflects the nature of AI-driven technological progress. In standard general-equilibrium setting, the rest of the economy adjusts endogenously to such a shock, preserving the factor-return gains of the progressing sector. We depart from this framework by introducing Finite Change, which permits a non-progressing industry to contract fully and exit equilibrium, thereby mimicking the forced redundancy of sectors displaced by AI-enabled technologies. While "small" technological shocks preserve the wage gains of the advancing sector in general equilibrium, we show that sufficiently "large" shocks-i.e., Finite Change-can overturn this outcome, leaving workers in the initially advancing sector worse off. Notably, this reversal can arise even when the initial technological improvement occurs exclusively in the labor-intensive industry, a result that cannot be obtained in models restricted to marginal (infinitesimal) changes.