<p>Since there is no clear understanding about whether and to what extent afforestation could affect grain production due to the squeeze it causes on arable land, taking the forestry carbon sinks (FCS) projects in pilot regions of China as a quasi-experiment, this paper evaluates the impact of the projects on grain production and its underlying mechanisms. The difference-in-differences method and the generalized synthetic control method are employed to address the potential selection biases originating from both the observables and the time-invariant unobservable. Based on panel data from the counties in Sichuan, we find that FCS projects have no significant impact on grain production. However, we do reveal that the FCS projects tend to reduce the area of the arable land devoted to grain production and to increase the number of agricultural labourers and the grain yield per hectare, while no significant impact on the mechanization of agricultural production and the use of fertilizers has been observed. We also demonstrate that the negative impact of FCS could be compensated by increasing the land production efficiency. Our findings provide the first evidence that FCS do not necessarily and negatively affect grain production and can even improve the efficiency of resource use.</p>