Nonlinear Effects of Digital Empowerment on Agricultural Green Total Factor Productivity in China

Achieving sustainable agricultural development requires both technological progress and well-designed public institutions. This paper examines whether regional digital development improves agricultural green total factor productivity (AGTFP) in China—an integrated measure of productivity, resource conservation, and environmental performance—and whether government fiscal intervention conditions that relationship. A city-level digital empowerment index is constructed from digital infrastructure, digital applications, and digital human capital, and AGTFP is measured with a Super-SBM model that incorporates undesirable outputs such as agricultural carbon emissions and excess fertilizer input. Using panel data for 226 prefecture-level cities over 2011–2020, two-way fixed-effects estimates show that digital empowerment is positively and significantly associated with AGTFP: a one-standard-deviation increase in digital empowerment corresponds to roughly 0.03 standard deviations of AGTFP, suggesting that digitalization can advance resource-saving and low-carbon agricultural production. Crucially, this sustainability dividend is not automatic. The positive association weakens as fiscal intervention intensifies, and threshold estimates identify a structural break at a fiscal-intervention ratio of about 0.29, above which the association turns negative—a pattern consistent with fiscal crowding out of market-led digital investment. These findings imply that digital dividends for sustainable agriculture depend on how public spending is allocated, not merely on how much is spent. This paper contributes city-level evidence on the boundary conditions of digital empowerment in agriculture and shows why linear, quadratic, and threshold specifications should be interpreted as distinct forms of nonlinearity rather than as a single policy optimum.