A Meta-Analysis of Distributional Treatment Effects in the Microcredit Literature

This study will develop new methodology for meta-analyses of distributional treatment effects in order to produce a new meta-analysis of microcredit interventions. Bayesian hierarchical models provide the framework for aggregation of quantile treatment effects and variance treatment effects, allowing for heterogeneous effects across studies while also estimating a generalized effect. I will also develop accompanying metrics of external validity, by extending the existing Bayesian pooling metrics to assess the heterogeneity in distributional effects across sites. I will consider a variety of modeling choices in order to derive robust and flexible hierarchical models and evaluate their relative performance using Monte Carlo simulation studies where appropriate. The resulting analysis should reveal the full distributional impact of microcredit access, and thus inform future policy decisions regarding microfinance institutions.

本研究旨在开发新的元分析方法，以综合分析分布性治疗效果，进而对微型信贷干预措施进行新的元分析。贝叶斯分层模型构成了聚合分位数治疗效果和方差治疗效果的框架，它允许在不同研究之间存在异质性效果的同时，估算出广义效应。我还将开发伴随的外部有效性指标，通过扩展现有的贝叶斯池化指标来评估不同地点间分布性效果的异质性。我将考虑多种建模选择，以推导出稳健且灵活的分层模型，并利用蒙特卡洛模拟研究对其相对性能进行评估。所得分析应揭示微型信贷获取的完整分布性影响，从而为未来关于微型金融机构的政策决策提供信息。