中文摘要
在许多情境中,例如基于地域的政策评估,潜在结果不仅取决于个体自身的处理状态,还取决于其邻居的处理状态。尽管如此,"双重差分"(DID)类估计量通常忽略了邻居之间的这种干扰。本文表明,在存在邻里干扰的情况下,经典 DID 估计量一般无法识别感兴趣的因果效应。为将干扰结构纳入 DID 估计,我在一个修正的平行趋势假设下,提出了处理组直接平均处理效应的双向稳健(doubly robust)估计量。我允许空间相关的分配机制,并在后续放宽了文献中的常见限制,如仅限直接邻居的干扰以及溢出函数设定正确的要求。此外,本文采用有限总体条件推断框架,并基于所提估计量的渐近分布讨论稳健推断。
Abstract
In many scenarios, such as the evaluation of place-based policies, potential outcomes depend not only on the unit’s own treatment but also on its neighbors’ treatments. Despite this, “difference-in-differences” (DID) type estimators typically ignore such interference among neighbors. I show in this paper that the canonical DID estimators generally fail to identify interesting causal effects in the presence of neighborhood interference. To incorporate interference structure into DID estimation, I propose doubly robust estimators for the direct average treatment effect on the treated under a modified parallel trends assumption. I allow spatially correlated assignments, and later relax common restrictions in the literature, such as immediate neighborhood interference and correctly specified spillover functions. Moreover, this paper focuses on a finite-population conditional-inference framework. Robust inference is discussed based on the asymptotic distribution of the proposed estimators.