Research:

Working papers

From Entry to Leadership: Early-Career Rotation and Leadership in Pakistan’s Civil Service

The Policy Preferences of Senior Civil Servants

(with Sultan Mehmood , Shachar Kariv and Ray Fisman)

Submitted Version: Oct 21, 2025.

Abstract. Policymakers balance priorities like poverty reduction, infrastructure development, healthcare, and education. We examine how nearly 1,000 high-level Pakistani officials across federal and regional governments navigate these trade-offs through a budget line experiment, allocating resources between their top two policy priorities (selected from nine) while varying the relative price of reallocation. Choices are generally consistent with maximizing a policymaking (constant elasticity of substitution) objective function. While there is considerable heterogeneity in policy prioritization—ranging from perfect complements to perfect substitutes—subjects most commonly view their top priorities as complements rather than substitutes. Certain individual attributes – such as gender and educational background – predict policy preferences in largely expected ways.

Improving Governance through Citizen Feedback Technologies: Evidence from Pakistan

(with Silvia Vannutelli and Sultan Mehmood)

First Draft Version: Feb 27, 2026.

Abstract. This paper explores how citizen feedback and bureaucratic incentives shape government performance in developing countries, drawing on administrative data from nearly 1.9 million citizen complaints in Pakistan. The data allows us to uniquely observe both citizen and official perspectives on public service delivery across districts and over time. Our findings reveal three key insights. First, the use of the digital platform varies significantly by geography and demographics, with women, older citizens, and less-educated individuals notably underrepresented. Second, there is a persistent 17-percentage-point gap between government-reported and citizen-perceived resolution rates, suggesting discrepancies in how outcomes are assessed. Third, bureaucratic transfers increase complaint resolution by about 6 percentage points, particularly in departments that provide tangible services. These results indicate that personnel mobility can serve as an effective accountability mechanism within bureaucratic systems, offering a direct and observable measure of citizen satisfaction that would otherwise be difficult to capture at scale.

AI Education as State Capacity: Experimental Evidence from Pakistan

(with Sultan Mehmood and Daniel Chen)

Revise and Resubmit, Journal of Development Economics

Training Policymakers in Econometrics

(with Sultan Mehmood and Daniel Chen)

Reject and Resubmit, The Economic Journal