Abstract

Investment in digital technologies can improve coordination and resource use while imposing substantial costs before complementary organizational capabilities develop. This study examines that tension using 8,337 observations from 1,599 Chinese listed enterprises over 2007–2020. Digital transformation is measured by the carrying value of information and communications hardware and software relative to total assets. Total factor productivity is estimated through the Olley–Pakes approach, with Levinsohn Petrin estimates used in a robustness exercise. The reported regressions identify a negative linear coefficient and a positive quadratic coefficient, consistent with a U-shaped relationship within the observed investment range. The fully controlled specification places the estimated turning point at 0.1373, compared with a sample mean digital investment ratio of 0.0090. The curvature persists under alternative fixed effects, a different productivity estimator, the exclusion of real estate firms, and additional city controls. Mechanism regressions suggest favorable contributions through working capital turnover and the educational composition of employment, alongside an adverse financial-allocation pathway associated with the financing constraint indicator. The turnover evidence is less decisive than the human capital evidence because the investment-to-turnover coefficient is not conventionally significant. Subsample estimates retain the same coefficient sign pattern across asset-size and ownership categories, although they do not establish identical effect magnitudes. These findings favor treating digital transformation as an organizational investment process whose productivity implications depend on complementary skills, operational integration, and the allocation of finance. The estimated turning point describes conditional marginal productivity within the fitted model; it is neither a universal spending prescription nor direct evidence that every enterprise follows the same temporal adjustment path.