比较案例研究中处理单元微观数据的异质性政策效应
Heterogeneous Policy Effects in Comparative Case Studies with Treated-Unit Microdata
- Karlstad Business School（卡尔斯特德商学院）
- Center for Societal Risk Research, Karlstad University（卡尔斯特德大学社会风险研究中心）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文提出一个识别框架，在比较单元缺乏个体数据时，结合处理管辖区内的双重差分对比与总体面板数据识别的总体平均效应，以识别异质性政策效应。
AI中文摘要：
政策改革有时伴随着实施管辖区内详细的个体层面数据，而潜在的比较管辖区仅能获得总体结果。本文开发了一个识别框架，用于在比较单元无法获得个体层面数据时识别异质性政策效应。该框架将处理管辖区内双重差分比较的处理效应对比，与从总体面板数据中识别的相容总体平均效应相结合。识别要求处理管辖区内存在相对平行趋势，同时满足从总体面板识别总体平均效应所需的假设。管辖区内的组成部分可以通过具有单一预处理期的重复横截面数据来估计。
英文摘要：
Policy reforms are sometimes accompanied by detailed individual-level data in the implementing jurisdiction, while only aggregate outcomes are available for potential comparison jurisdictions. This article develops an identification framework for heterogeneous policy effects when individual-level data are unavailable for the comparison units. The framework combines treatment-effect contrasts from difference-in-differences comparisons within the treated jurisdiction with a compatible population-average effect identified from aggregate panel data. Identification requires relative parallel trends within the treated jurisdiction together with the assumptions needed to identify the population-average effect from the aggregate panel. The within-jurisdiction component can be estimated from repeated cross-sections with a single pretreatment period.
