The Unintended Effects of Bottom-Up Accountability: Evidence from a Field Experiment in Peru

Past research suggests that increasing citizen political knowledge and coordination can improve government performance via “bottom-up accountability,” where mobilized local communities exert pressure on elected officials through democratic processes. A randomized field experiment in Peru demonstrates that interventions to promote bottom-up accountability can sometimes have unintended effects on government performance, among other outcomes. I find that accountability workshops reduce participation in the district’s “participatory budgeting” process and increase support for civil unrest as a tool for sanctioning politicians. Although the intervention increases the initiation of recalls for poor-performing mayors, these mayors respond to the recall threat by further reducing their effort. Taken together the evidence suggests that improved information and coordination of local elites is not sufficient to improve government performance where it has previously lagged and can in fact be counterproductive. ∗Postdoctoral Fellow, Niehaus Center, Princeton University; email: rsexton@princeton.edu. The intervention was designed and implemented in collaboration with Innovations for Poverty Action (Lima) and the Centro de Investigación y Promoción del Campesinado (CIPCA), a member of the Propuesta Ciudadana network. I thank Maria Luisa Zeta, Gonzalo Manrique and Juan Manuel Hernandez-Agramonte of IPA for their research assistance, as well as Epifanio Baca and Gustavo Avila of Propuesta Ciudadana, and gratefully acknowledge financial support for this research from the Governance Initiative of the Abdul Latif Jameel Poverty Action Lab at MIT. Thanks for helpful comments from Matthew Bird, Graeme Blair, Darin Christensen, Mike Findley, Mike Gilligan, Kosuke Imai, Livio di Lonardo, Mai Nguyen, Tom Pepinsky, Cyrus Samii, Jake Shapiro, and seminar participants at APSA, PELA and Polmeth. This research comes under IPA IRB Protocol no. 13696.