Digital-Intelligent Technology Application and Green Technological Output: Production-Frontier Shortfall and Provincial Environmental-Policy Context in Chinese Manufacturing Firms

Digital-intelligent technologies are increasingly reshaping how manufacturing firms organize innovative resources for sustainable development. Using 26,513 firm-year observations from Chinese A-share listed manufacturing firms during 2011–2023, this study examines whether and under what institutional conditions disclosed digital-intelligent technology (DIT) application is associated with green technological output. Green technological output is measured as ln(1 + green patent applications). The proposed allocation-correction pathway is examined using a standardized signed production-frontier shortfall—the distance between firm-level Levinsohn–Petrin total factor productivity and the industry-year 99th-percentile benchmark—as a relative allocation-efficiency proxy rather than a structural measure of factor-market misallocation. Firm and year fixed-effects estimates show that DIT application is positively associated with green technological output and negatively associated with the frontier shortfall; a larger shortfall is negatively associated with green technological output. A descriptive product-of-coefficients decomposition is small and is not interpreted as causal mediation. Higher values of the provincial environmental-policy proxy (ER) are associated with a weaker conditional DIT–green-output relationship and a stronger negative DIT–shortfall relationship in the p99 specification, although the latter interaction is sensitive to the frontier threshold. The supplementary instrumental-variable estimate is imprecise, and its weak-instrument-robust confidence set includes zero. The findings therefore describe conditional associations and suggest that the sustainability value of disclosed DIT application depends partly on firm resource-use conditions and the provincial environmental-policy context.