预测干部容错:效率考量、合法性压力与领导特征——基于机器学习方法的预测建模
Predicting the Diffusion Logic of the Cadre Fault-tolerant Mechanism:Efficiency,Legitimacy,and Leadership——Predictive Modeling Based on Machine Learning
刘展余 1余洋 2雷玉琼2
作者信息
- 1. 复旦大学国际关系与公共事务学院
- 2. 湖南大学公共管理学院
- 折叠
摘要
国家对于干部容错机制的诉求愈发强烈,但是近年来该政策在国内的扩散速度逐步放缓.要解决上述问题,关键在于研究干部容错机制的扩散逻辑.在政策创新扩散逻辑相关研究中,既有研究方法大多遵循"样本内解释"的思路.这使得相关研究结论的科学性,尤其是泛化能力,难以得到保证.为突破这一局限,本文采用"样本外预测"的思路来探索干部容错机制的扩散逻辑.具体来看,本文基于政策创新扩散理论构建了分析框架,使用机器学习方法训练了地方政府采纳干部容错机制的预测模型,并在确保模型预测性能的前提下呈现干部容错机制的扩散逻辑.模型解读结果表明,干部容错机制的扩散主要由行动者逻辑主导,其次是效率逻辑,最后是合法性逻辑;效率逻辑、合法性逻辑及行动者逻辑三个维度中,最具影响力的特征分别为治理规模、同级采纳和主官任期.地方政府采纳干部容错机制的概率与治理规模具有负向关系,与同级采纳具有正向关系,与主官任期具有倒U形关系.为推进干部容错机制有序发展,本文建议地方政府强化机制扩散中的领导驱动效应、科学规划干部容错机制扩散路径.
Abstract
The national government has been increasingly vocal about the cadre fault-tolerant mechanism(CFTM),however,the spread of this policy in China has gradually slowed down in re-cent years.To solve the above problems,the key is to study the diffusion logic of CFTM.In the re-search on the diffusion logic of policy innovation,most research methods generally follow the"with-in-sample explanation"approach.This makes it difficult to ensure the scientific validity,especially the generalizability,of the research conclusions.To overcome this limitation,this paper adopts the"out-of-sample prediction"approach to explore the diffusion logic of CFTM.Specifically,based on the policy innovation diffusion theory,this paper constructs an analytical framework,uses machine learn-ing methods to train a predictive model for local governments adopting CFTM,and presents the diffusion logic of CFTM while ensuring the predictive performance of the model.The model interpretation results re-veal that the diffusion of CFTM is mainly dominated by actor logic,followed by efficiency logic and legitima-cy logic.Among the three dimensions of efficiency logic,legitimacy logic,and actor logic,the most influen-tial features are governance scale,peer adoption,and the tenure of the top leader,respectively.The proba-bility of local governments adopting CFTM is negatively related to governance scale,positively related to peer adoption,and has an inverted U-shaped relationship with the tenure of the top leader.To promote the orderly development of CFTM,this paper suggests that local governments strengthen the leadership-driven effect in the diffusion of CFTM and scientifically plan the diffusion path of CFTM.
关键词
干部容错机制/政策创新扩散/机器学习/预测建模Key words
Cadre Fault-tolerant Mechanism/Policy Innovation and Diffusion/Machine Learning/Predictive Modeling引用本文复制引用
出版年
2024
NETL
NSTL
万方数据