基于OLS和分位数回归的农地整理管护绩效研究

摘要

This paper is aiming to evaluate the performance of supervision and maintenance after ownership transfer of rural land consolidation, and thereby to reveal the significant impact factors at different quantiles. In order to obtain the goal, Structure-Conduct-Performance paradigm has been adapted to probe into the theoretical cause and effect and thereafter OLS Regression and Quantile Regression have been successively employed to examine the data. The results show that 1)the mean value and standard deviation of performance is 35.620 and 5.583, while the performance is non-normal distributed overall and single-peak; 2)institutional structure and conducts affect performance of supervision and maintenance significantly, as Structure-Conduct-Performance paradigm predicts; 3)there exist various affecting factors when performance of supervision and maintenance falls into low-level, middle-level, or high-level interval, respectively. It is concluded that using quantiles could be helpful to analyze the distribution features of performance on supervision and maintenance and more importantly, comprehensive improvement measures should be integrated into each level interval.%研究目的：在对农地整理管护绩效进行测度的基础上，揭示管护绩效的影响因素及其在不同分位点处的变化特征。研究方法：结构—行为—绩效的研究范式和基于OLS回归、分位数回归的实证检验。研究结果：（1）实证区域管护绩效均值为35.620，标准差为5.583，呈非正态的“单峰模式”分布；（2）OLS回归结果指出，制度结构和管护行为显著影响农地整理管护绩效，验证了“结构—行为—绩效”分析框架的有效性；（3）分位数回归的结果说明，低、中、高水平区间内管护绩效的影响因素不完全一致。研究结论：分位数的研究视角有助于深入分析管护绩效的样本分布区间特征，在不同绩效水平区间应综合运用多种管理措施以提升管护绩效。