Addressing weak links in government implementation at scale: Experimental evidence from a school governance reform in Tanzania
Journal of Development Economics · 2026 · [{"name": "Jacobus Cilliers", "affiliation": ["Georgetown University"]}, {"name": "James Habyarimana", "affiliation": ["Georgetown University"]}]
中文摘要
本文展示了低成本创新如何改善成本高昂的政府项目的实施。我们以实验方法评估了坦桑尼亚一项新学校治理项目在全国范围的推广。作为该项目的组成部分，学校督学开展全校访问（WSV），针对学校领导层、教师和家长，形成一组针对具体学校、旨在改善办学质量的评级与建议。然而，学校实施这些建议的激励较弱，而负责开展后续走访的地方监督员隶属于另一个部委，无法直接获取全校访问报告。为应对这一难题，在随机抽取的部分项目学校中，监督员收到短信，被告知主要建议，并被鼓励跟进学校以确保落实。在接受全校访问的学校样本中，两年后短信使学生学习成绩提高了0.11个标准差。对于在各个领域获得较差督学评级的学校，处理效应更大，这与监督员依据量身定制而非一般性建议采取行动相一致。所观察到的收益集中于被某捐助方项目覆盖的地区，该项目为地方官僚提供了用于监督的资源。我们的发现表明，破解实施环节中的具体瓶颈，可显著提升政府项目在大规模推广中的成效。
Abstract
This paper demonstrates how low-cost innovations can improve the implementation of costly government programs. We experimentally evaluate the nationwide roll-out of a new school governance program in Tanzania. As part of the program, school inspectors conduct Whole School Visits (WSVs) producing a set of school-specific ratings and recommendations—targeted at leadership, teachers, and parents—for improving school quality. However, schools faced weak incentives to implement the recommendations, and the local monitors responsible for conducting follow-up visits, who report to a different ministry, lacked direct access to WSV reports. To address this challenge, in a randomly sampled subset of program schools, monitors received text messages informing them of the main recommendations and encouraging them to follow up with schools to ensure compliance. Within the sample of schools that received WSVs, text messages improved student learning by 0.11 standard deviations after two years. Treatment effects are larger for schools that received worse inspection ratings in each domain, consistent with monitors acting on tailored rather than generic recommendations. Observed gains are concentrated in regions exposed to a donor program that provided local bureaucrats with resources to monitor. Our findings show how addressing specific bottlenecks in implementation can significantly enhance the effectiveness of government programs at scale.