Analysis of the Impact of Fiscal Revenue and Expenditure on China’s Grain Production Using Panel Double-Kink Regression Model

This study examines whether the relationship between provincial fiscal revenue and expenditure measures and grain production in China is nonlinear. Using a balanced panel of 31 provinces from 2007 to 2021, we analyze major revenue-side and expenditure-side fiscal instruments, including the cultivated land occupation tax, value-added tax, agricultural insurance subsidies, agricultural loan interest subsidies, rural minimum living security subsidies, education expenditure, and transportation infrastructure expenditure. To identify regime-dependent changes in estimated associations, we employ panel kink and double-kink regression models with endogenously estimated kink points. The results suggest that the estimated relationships are intensity-dependent rather than constant. The cultivated land occupation tax exhibits a kinked relationship with grain production, with a more positive association beyond a certain level. Agricultural insurance subsidies display a double-kink pattern, with the strongest positive estimated association concentrated in an intermediate range of the subsidy measure. Rural minimum living security subsidies are positively associated with grain production at lower levels, but this association weakens and may become negative after the estimated kink point. Overall, the findings suggest that the relationship between fiscal variables and grain production depends not only on policy direction but also on the levels of the fiscal measures.