Enterprise Digitalization and Estimated Carbon Exposure: Within-Firm Evidence and Sustainability Implications for Pollution-Intensive Industries

Using 21,879 firm–year observations from China’s A-share listed companies (2009–2021), we examine the within-firm association between enterprise digitalization and estimated carbon exposure constructed as industry–year carbon intensity multiplied by firm revenue. This measure captures sales-weighted exposure to an industry’s average carbon intensity; it is not observed firm-level greenhouse-gas emissions or environmental performance. Pooled OLS associates a one-standard-deviation increase in digitalization (SD = 1.204) with 0.1009 log points (approximately 10.1%) higher estimated carbon exposure, whereas the firm- and year-fixed-effects estimate is 0.0035 log points (approximately 0.35%) and statistically indistinguishable from zero. With an HDFE standard error of 0.0098, the corrected two-sided 5% minimum detectable effect for 80% power is 0.0275 in raw Digi units, equivalent to 0.0331 log points or 3.31% per SD. In heavy-pollution firms, the split-sample coefficient is positive (3.09% per SD), but the formal interaction test does not reject equality across pollution groups. These are estimator and sectoral patterns, not evidence of a causal rebound mechanism. For sustainability policy, digitalization indicators should not substitute for directly measured emissions, carbon pricing, or enforceable emissions constraints. The sample ends in 2021 and does not evaluate the subsequent operation of China’s national emissions trading system.