Working Papers

It Takes a Village Election: Turnover and Performance in Local Bureaucracies

(with Sam Bazzi, Masyhur Hilmy, Benjamin Marx, and Andreas Stegmann) (Working Papers: World Bank, NBER, CEPR) 

Abstract: In many countries, local governments struggle with inefficiency and corruption, often perpetuated by entrenched elites. This paper explores how leadership changes affect local bureaucratic performance. Combining personnel and citizen surveys with a regression discontinuity design in a large sample of Indonesian villages, we show that turnovers in village elections revitalize local bureaucracies, disrupt nepotistic networks, and improve local government performance. Bureaucrats serving new leaders are more engaged and less likely to be tied to past or present village officials, resulting in a more responsive bureaucracy that interacts more with citizens and better understands their needs. This improves public service provision, measured in both administrative data and citizen surveys. Overall, our results show that leadership changes can mitigate elite capture and improve governance at the grassroots level.

Horizontal Equity of Taxation: Citizen Beliefs and Policy Preferences (with Pierre Bachas, Christopher Hoy, and Anders Jensen) (Working Paper: World Bank) 

Abstract: Horizontal inequity occurs when employees and self-employed with the same income end up with different effective tax burdens, due to the difficulty of enforcing taxes on self-employed. Based on detailed micro-tax simulations models integrated with household surveys in 25 developing countries, we show that tax systems incur large horizontal inequities in practice and that reforms which improve vertical equity worsen horizontal equity by a comparable amount. An in-person survey in Pakistan and online surveys across multiple countries reveal widespread concern about horizontal equity. Randomized information treatments heighten this concern but do not shift tax preferences toward addressing horizontal inequity. 

How Do Organizations Learn? The Diffusion of Scientific Evidence on Generative AI  (with Andreas Stegmann and Mattie Toma) (Working Paper: World Bank, Summary: VoxDev) 

Revise and Resubmit, Journal of Political Economy: Microeconomics

Abstract: We study how scientific evidence diffuses through a framed field experiment at the World Bank. We experimentally vary (i) whether research findings on the impacts of generative AI are shared with senior or junior staff, and (ii) beliefs about peers’ adoption of and attitudes toward the evidence. Providing evidence to senior staff significantly increases evidence transmission, as measured by self-reported sharing, engagement with study materials, and colleagues’ recall of study details. In contrast, shifting second-order beliefs about peers’ adoption of and attitudes toward the evidence has no detectable effect. The results highlight the importance of organizational hierarchy in evidence dissemination.

Rebuilding the Social Compact (with Adnan Khan, Asim Khwaja, and Ben Olken) (AEA Registry Link, Short-run impacts PDF)

Abstract:  The social compact between citizen and state – whereby a citizen pays taxes and receives public goods and services – is a critical link in the development process.  In many countries, this compact is broken: citizens do not receive high quality services, because resources are limited by low levels of tax revenue. The low quality of services further leads to a low willingness to pay taxes in addition to a broader lack of trust in the state.  This paper examines whether strengthening the link between property taxes and services can increase citizens’ willingness to pay taxes, improve service delivery, and ultimately revitalize the social compact through a large-scale, long-term randomized controlled trial implemented with the provincial government of Punjab, Pakistan.  We test four interventions: “Ask,” where tax collectors elicit citizen preferences for local services; “Deliver,” where property tax revenue is earmarked for and used to provide local services; “Deliver and Tell,” where service delivery is paired with information to citizens; and “Ask, Deliver, Tell,” which combines preference elicitation, service delivery, and communication to citizens.  We estimate positive but modest effects on tax payments and attitudes toward the state, despite successful delivery of services in treated localities. Effects are largest in the “Deliver and Tell” arm, underscoring the importance of making tax–benefit linkages both salient and credible. 

Is There a Gender Gap In Taxation? (with Hitomi Komatsu and Ceren Ozer)

Revise and Resubmit, World Bank Research Observer 

Abstract:  Most tax systems do not explicitly distinguish between women and men, yet they may generate different outcomes across genders. In low- and middle-income countries, gender disparities in income, asset ownership, economic activities, and interactions with tax authorities shape individuals’ exposure to taxes, compliance burdens, and experiences with tax administration. This paper reviews the emerging literature on gender and taxation in developing countries and organizes the evidence around four channels through which gender differences arise: tax policy and administration, underlying socioeconomic inequalities, access to information and networks, and behavioral responses to taxation. The evidence suggests that ostensibly gender-neutral tax systems can produce gender-differentiated outcomes in practice. We discuss the implications of these differences for both equity and efficiency in tax design and administration, highlight the limitations of the existing evidence base, and identify priorities for future research and policy.

Electoral Competition and Politician Behavior (PDF)

Publications

Building Trust in in the State with Information: Evidence from Urban Punjab (with Adnan Khan, Sanval Nasim, and Andreas Stegmann) 

Journal of Public Economics, October 2021, Volume 202, 104494.

Abstract: Can communication designed to increase support for government policy and shift perceptions of state capacity redress deep-rooted mistrust in state institutions? This paper finds providing information on past state effectiveness, highlighting citizens’ cooperation in enabling past effectiveness or appealing to religious authorities’ support for government policy have limited impact on support for policy, perceptions of state capacity and trust in the state in Pakistan. This holds true on average and across important dimensions of heterogeneity after comparing treatment effects to those induced by an experimenter demand treatment. This paper highlights the limits of using information to build trust in state institutions, and the importance of measuring experimenter demand.

Selected Works in Progress

Improving Property Tax Collection with Computer Vision (with Sher Afghan Asad, Adnan Khan, Ben Olken, and Jie Zhou) 

Abstract:  Economic growth in developing countries is often limited by the state’s inability to raise tax revenue.  In many countries, tax administration systems rely on infrequently updated and out-of-date property tax valuations, and tax officials often employ significant discretion when assessing properties. These factors can lead to errors that could increase tax leakages or lower citizen trust in the state. This study addresses this challenge in two steps: first, by developing a computer vision algorithm that can use property images to predict property assessments and second, by testing how well the algorithm performs in identifying properties for reassessment.