Abstract
Does digital infrastructure automatically generate productivity dividends, or do institutional conditions shape its returns? This paper examines broadband penetration and manufacturing labor productivity using panel data from 30 Chinese provinces over 2012–2023. Fixed-effects estimation yields a significant positive effect (β = 0.89, p < 0.05). A Hansen panel threshold test identifies a significant threshold at marketization index 7.31 (p < 0.001), but a Wald test reveals the cross-regime coefficient difference is statistically insignificant (p = 0.60). A continuous interaction model fits better (ΔAIC = 14.76), with a positive interaction term (β = 0.28, p < 0.01), indicating that marketization functions as a gradual amplifier rather than a binary switch. Exploiting the staggered rollout of the ”Broadband China” demonstration city policy as a quasi-natural experiment, event study confirms parallel pre-trends (joint F-test: p = 0.907). While the average treatment effect is statistically insignificant, heterogeneous DID reveals that the policy’s effect on industrial output is strongly moderated by marketization (interaction β = 751.56, p < 0.01). Mechanism analysis identifies technological innovation as the primary transmission channel (β = 165.02, p < 0.001). The evidence implies that institutional reform should accompany digital infrastructure investment to maximize industrial productivity gains—though in provinces with weaker institutions, broadband investment requires careful governance to avoid counterproductive outcomes.

This publication has 25 references indexed in Scilit: