基层“微腐败”数字痕迹监管的时空演化规律方法论探索

摘要： 本研究聚焦于数字治理背景下基层“微腐败“的行为演化规律，试图为微腐败大数据时空痕迹研究者提供一种全新的方法论，通过构建乡镇的多源数字痕迹面板数据集，综合运用空间计量经济学与时间序列分析方法，系统揭示基层微腐败数字痕迹的时空分异与耦合演化机制。研究发现：第一，微腐败风险呈现显著的空间正自相关性（全局Moran's I = 0.324，p < 0.01），形成稳定的“高-高”与“低-低”集聚区。其中，经济发达地区多表现为高频小额、跨平台操作的“技术规避型”痕迹，而经济欠发达地区则呈现低频集中、单平台操作的“传统粗放型”特征。第二，微腐败数字痕迹存在明显的季节性周期波动，STL分解结果显示，春节（t = 4.32）与年末（t = 3.89）等关键制度节点异常值显著上升，且发达地区的季节性峰值更为隐蔽、周期更长。结论表明，基层微腐败的数字化演化已形成“空间聚类—季节适配”的双重逻辑，其行为模式与区域发展水平、数字治理能力深度耦合。建议监督策略应从“一刀切”转向“精准制导”，兼顾区域差异化特征与周期动态性变化，构建“时空智能预警”治理框架，并关注数字化监督可能引发的风险“挤出”与策略“演化”效应。

Abstract: This study focuses on the behavioral evolution patterns of grassroots“micro-corruption” within the context of digital governance. By constructing a multi-source digital trace panel dataset encompassing 200 townships in the Yangtze River Delta region from 2018 to 2023, and employing an integrated methodology of spatial econometrics and time-series analysis, it systematically reveals the spatiotemporal differentiation and coupled evolutionary mechanisms of digital traces associated with grassroots micro-corruption. The findings are twofold. First, micro-corruption risk exhibits significant positive spatial autocorrelation (Global Moran's I = 0.324, p< 0.01), forming stable "high-high" and“low-low” agglomeration zones. Within these zones, economically developed areas predominantly display traces characterized as the“"technology-evasion type,” featuring high-frequency, small-amount transactions and cross-platform operations. In contrast, less developed area