Non-grain production on cultivated land and grain production resilience: evidence from 291 prefecture-level cities in China

Shifting cultivated land from rice, wheat, and maize to other crops reduces the area available for staple-grain cultivation and leads to the reallocation of agricultural production factors and services. Using a 2004–2024 panel of 291 Chinese prefecture-level cities, we match annual land-cover data with remote-sensing maps of the three staple crops to measure the non-grain production rate (NGR) on cultivated land. Grain production resilience (GPR) is assessed across three dimensions: resistance capacity, recovery capacity, and adaptive capacity. Double machine learning (DML) estimates show that a 10-percentage-point increase in NGR lowers GPR by 0.00261 on average, equivalent to about 4.1% of the sample mean. All three components decline significantly. The negative effect also appears in the alternative-index and Bartik instrumental-variable estimates. A higher NGR is associated with significantly fewer granted agricultural patents, suggesting that weaker agricultural technology innovation may be one channel linking NGR to lower GPR. The adverse effect of NGR on GPR is larger at higher levels of economic development and increases further as the economic development level passes the two estimated thresholds. Bootstrap tests indicate larger negative effects in low-slope areas and areas with larger-scale agricultural operations. Cultivated-land protection should extend beyond maintaining the total area of cultivated land to preserving space for staple-grain cultivation and the technologies, services, and production organization that support it. Policies should be differentiated according to the stage of economic development and local grain-production conditions.