Abstract: At present, farmland abandonment is a widespread problem in many poor mountainous areas of China. This type of land use change in sensitive environment has positive impact on the ecological security of these areas, but affects negatively on grain production and food security. For national food security reasons, large amounts of basic farmland in mountainous areas are still required. Therefore the contradiction between ecological security and food security is hard to resolve. Farmland transfer, however, may offer a solution to this problem. Understanding the characteristics and determinants of farmland transfer- in provides insights for decision making on the rational distribution of basic farmland and grain for green projects, but existing research is insufficient for mountainous areas. Based on a field survey of 1015 households in the "two wings" of Chongqing using participatory rural appraisal (PRA) tool, this study analyzes the basic characteristics of farmland transfer- in at plot level. It quantitatively analyzes the influencing factors on the farmland area that farming households transfer in by using the Tobit model. The results show that: (1) As free-of-compensation transfer is common in the study area, the surveyed households are unwilling to transfer out large plots located close to residence. However, farming households who transfer in farmland tend to choose plots with relatively good quality or convenient for farming, and have a low demand for plots with relatively poor production conditions. Thus these plots are more likely to be abandoned. (2) The farming households who transfer in farmland are often small farmers with low cropping income and their plots are generally small. With regard to farmland management, the surveyed farming households tend to grow cash crops on the plots that they transfer in, while labor and fertilizer input to these plots are lower than that of own plots. (3) The Tobit model analysis suggests that farm labor force, household farming income, quantity of farming equipment, and cultivated area have significant positive impacts on the farmland area that households transfer in. On the other hand, age of household head, average plot area, and Wulong County have significant negative impacts. Other variables, such as off-farm income, quantity of household livestock, distance from market, road, and Wushan County, have no significant impact on the farmland area that farmers transfer in. (4) In order to foster land transfer market in the mountainous areas, targeted measures need to be adopted by the government, such as developing land consolidation projects, adjusting grain production subsidies and cropping structure, increasing the amount of agricultural machinery purchase allowance, and so on. This study also identified issues that require further research, for example, why free-of-compensation transfer is common in the mountainous areas and what is the role of the government in farmland transfer.

Key words: Chongqing, determinants, farmland transfer-in, mountainous area, plot, Tobit model