贫困-心理健康梯度的出现：来自 2005 年巴基斯坦大地震的证据
The emergence of the poverty-mental health gradient: Evidence from the great Pakistan earthquake of 2005
Journal of Development Economics · 2026 · [{"name": "Tahir Andrabi", "affiliation": ["Pomona College"]}, {"name": "Benjamin Daniels", "affiliation": ["Harvard Global Health Institute"]}, {"name": "Jishnu Das", "affiliation": ["Georgetown University"]}]
中文摘要
贫困—心理健康梯度在不同人群之间存在显著差异，其程度远超单纯抽样变异所能预测。我们将这种异质性作为主要的实证研究对象。利用2005年巴基斯坦大地震四年后、从距活动断层线0–75公里范围内126个村庄收集的数据，我们发现，在未受影响的人群中，消费贫困与心理健康之间不存在相关性。相比之下，在最靠近断层的人群中，消费每增加一个对数单位，心理健康改善0.12个标准差；而在20公里之外，这一关系精确为零。该梯度对替代性距离阈值、样本划分和二元心理健康测度均保持稳健，且并非由急性家庭创伤所中介。这些结果表明，大规模逆境可能激活贫困—心理健康梯度，从而有助于解释更广泛文献中的异质性。
Abstract
The poverty–mental health gradient varies substantially across populations, far more than sampling variation alone would predict. We treat this heterogeneity as a primary object of empirical interest. Using data collected four years after the 2005 Great Pakistan Earthquake from 126 villages spanning 0–75 km from the activated fault line, we show that in unaffected populations there is no correlation between consumption poverty and mental health. By contrast, among those nearest the fault, mental health improves by 0.12 standard deviations per log unit of consumption, while the relationship is a precise zero beyond 20 km. This gradient is robust to alternative distance thresholds, sample splits, and binary mental health measures, and is not mediated by acute household trauma. These results suggest that mass adversity can activate the poverty–mental health gradient, helping explain heterogeneity in the broader literature.