智能化转型、技能适配与就业结构分化

Abstract

Amidst intelligent transformation reshaping production systems, this study employs an “educational orientation-skill alignment” perspective to examine how the education system shapes and matches workforce skills. Integrating China Family Panel Studies (CFPS) data to measure industrial over-education rates and matching it with data from Chinese A-share listed companies, this study identifies skill adaptation conditions in the labor market and empirically tests its moderating effect on employment during intelligent transformation. The results suggest that intelligent transformation fosters overall growth in labor demand through human-machine complementarity. Structurally, however, a hierarchical differentiation emerges: It exhibits complementary effects on high-skilled labor while presenting substitution effects on low-skilled labor. The realization of these employment effects is critically moderated by over-education. On one hand, under conditions of over-education, workers face a triple dilemma of skill redundancy trap, “degree signaling failure”, and degraded adaptive capacity, leading to ineffective skill synergy and thereby suppressing the complementary effects of intelligent transformation on both high-skilled and overall labor. On the other hand, over-education enhances the overall cognitive level of the labor force through credential accumulation, exerting a human capital buffering effect that effectively mitigates the intelligent transformation’s substitution effect on low-skilled labor. This study concludes that during the employment structure transformation in the intelligent era, the “human capital buffer” formed by over-education exhibits differentiated efficacy. In the short term, highly educated individuals leverage their foundational cognitive and learning abilities to downwardly adapt, utilizing humans’ comparative advantage in non-routine tasks to slow the displacement of low-skilled positions. However, when intelligent technologies extend into high-skill fields requiring deep collaboration, generalized academic credentials often fail to achieve precise skill alignment. Instead, they may undermine the complementary effects of deep human-machine collaboration, thereby constraining the upgrading of the overall employment structure and impeding the unleashing of technological dividends. Accordingly, this study proposes the following policy implications: enhancing the flexibility of the education system to curb skill redundancy; promoting the establishment of an ability-oriented signaling mechanism in the labor market; implementing differentiated intervention strategies to foster deep human-machine collaboration in high-skilled fields; and strengthening social security and reskilling programs for vulnerable groups to promote technology for good and inclusive development.