交易费用、农户契约选择与林地规模经营模式研究：驱动力、效率与机制设计

农户林地经营效率是衡量集体林权制度改革成效的重要经济技术指标，农户林地规模化是提高林地经营效率的重要路径和发展趋势，农户林地规模经营模式的选择是解决林地经营规模化的关键问题。本项目以农户行为、交易成本、契约效率等相关理论为指导，建立农户林地规模经营模式选择及其效率的理论分析框架、关键因素指标体系及定量分析模型。以江西省为研究对象，运用Multinomial Logit回归、BOOTSTRAP分位数回归模型探索交易费用、地块特征和农户特征对农户林地规模经营模式中契约方式和对象选择的影响机理；利用三阶段DEA-Tobit方法，从统一的框架下比较不同林地规模经营模式的农户契约效率差异；继而研究契约方式与林地规模经营模式的匹配策略。在上述基础上，提出激励农户提高林地经营效率、进而提高农户收入水平的政策建议，为决策部门制定和完善后续支林惠农配套政策提供科学依据，具有较大的理论价值和现实指导意义。

Farmers woodland management efficiency is an important economic indicator to measure the effectiveness of the Reform of Collective Forest Property Right System. Farmers woodland scaling is an important path and trends to improve woodland management efficiency. It is a key issue to research farmer choosing the woodland scale management modes and its efficiencies. Taking farmer behavior theory, transaction costs theory, contract efficiency theory as guides, the project will establish the theoretical framework of farmer selecting woodland scale management mode and it efficiency, a key factor index system and quantitative analysis model. Taking Jiangxi province as an object of study, it will use Multinomial Logit Regression and BOOTSTRAP Quantile Regression to explore the influence Mechanism that transaction costs, woodland block and household characteristics influence on farms selection behaviors of contract ways and contract objects in woodland scale management modes, use three-stage DEA-Tobit analysis method to compare the difference on farmers contract efficiencies of each woodland scale management mode from a united framework. And then study the matching policies of the woodland scale management modes and the contract ways. Based on the above, it will propose policy recommendations that incentives farmers to improve land management efficiency and increase the income levels of forestland, and provide a scientific basis for decision-making to formulate and improve the follow-up policies for supporting farmer and benefiting farmer. It will have great theoretical value and practical guiding significance.