警察自主权、数据驱动策略与暴力：来自巴西警务改革的证据

Journal of Development Economics · 2025 · [{"name": "Rafael Parfitt", "affiliation": ["Fundação de Apoio à Universidade Federal de São Paulo"]}, {"name": "Bruno Pantaleão", "affiliation": ["Citrix (Switzerland)", "Federal Government of Brazil"]}, {"name": "Alberto L. W. Kopittke", "affiliation": ["Universidade Federal do Rio Grande do Sul"]}]

中文摘要

我们评估了巴西一项结果导向型管理项目的影响，该项目旨在通过更高效的资源配置、数据驱动的人员部署以及在高暴力地区实施定向干预，改善警察服务供给。利用该项目在二十年间于巴西九个州的2,355个市镇交错推行的实施安排，我们采用交错双重差分方法估计其对凶杀率的因果效应。我们发现，改革实施后凶杀率大幅下降。这一效应在弱势子群体中尤为明显，例如黑人群体和受教育程度较低者。此外，我们还发现，在基线暴力水平较高的市镇，凶杀率降幅更大，这表明结果导向型警务管理在暴力最严重的地区尤其有效。

Abstract

We evaluate the impact of a results-based management program in Brazil designed to improve police service delivery through more efficient resource allocation, data-driven personnel deployment, and targeted interventions in high-violence areas. Exploiting the staggered implementation of the program across 2,355 municipalities in nine Brazilian states over two decades, we estimate its causal effect on homicide rates using a staggered difference-in-differences approach. We find that the reform substantially reduced homicide rates following implementation. The effects are particularly pronounced among vulnerable subgroups, such as Black individuals and those with low education levels. Additionally, we document larger reductions in homicides in municipalities with higher baseline levels of violence, suggesting that results-based police management is especially effective in the most violent areas.