Ongoing Project on Stephen D. O'Connell

Economic Recovery and Social Cohesion: A Field Experiment with Capital Grants in Post-Conflict Iraq

Mon, 01 Jun 2026 00:00:00 +0000

How does the uneven distribution of economic recovery affect social cohesion after war? While new economic opportunities can bind communities together, inequality may generate resentment among those left behind. We study a $2,000 capital grant program in post-conflict Iraq, measuring economic outcomes and seven dimensions of social cohesion. We designed a randomized controlled trial to identify the direct effects of the grants among beneficiaries and indirect effects on their first-degree social ties and professional peers. Beneficiaries experience large economic gains that strengthen over time (a 24.6 percentage point increase in business ownership and 0.81 SD higher business revenues by twelve months), alongside improved trust and reduced grievances and perceived competition toward their community. Early increases in trade, debt, and transfers between beneficiaries and their professional peers, consistent with local sharing norms, fade by twelve months. Social ties and professional peers report persistent directed grievance toward the specific beneficiary with whom they are connected (−0.21 SD for professional peers, −0.18 SD for social ties; q < 0.01 at both six and twelve months), but no changes in directed trust and contact allow us to rule out an overall deterioration in dyadic relationships. Their general attitudes and behaviors toward the broader community, including trust, civic engagement, and inclusion, are unchanged across both waves, with 95% confidence intervals excluding indirect exposure effects larger than 0.20 SD. These results show that the social cost of uneven recovery is concentrated in directed, tie-level perceptions rather than in non-recipients’ general attitudes.

ML-based geographic sampling frames miss transitory populations in fragile regions

Fri, 01 May 2026 00:00:00 +0000

Survey research on conflict and displacement depends on the reliability of sampling frames. We show that common approaches to developing these frames may underrepresent populations central to this research: displaced people, returnees, and civilians exposed to violence. We develop a hybrid approach to sampling frame generation that combines machine-learning (ML)-generated building footprints with satellite imagery. We test the approach in displacement-affected Iraqi communities. Our approach achieved 87% residential accuracy overall, and our data reveal non-random omissions in the common ML-only approach. We find systematic coverage differences across frame-generation methods. Manual methods capture more rural internally displaced persons and urban returnees, due to informal shelters and wartime reconstruction. Our hybrid ML-and-satellite sampling can mitigate coverage error and improve inference about conflict and displacement.

Testing Threshold Criteria for Return Migration: A Survey with Forcibly Displaced Venezuelans

Fri, 20 Feb 2026 00:00:00 +0000

Standard economic models treat migration as a human-capital investment driven mainly by economic returns. Migration under forced displacement is a different decision problem, often governed by whether conditions in the home country cross a minimum threshold of improvement rather than by compensatory tradeoffs across attributes. This project develops and tests a threshold model of return migration, combining classic migration theory with behavioral heuristics of threshold decision-making. Building on qualitative fieldwork with Venezuelan refugees in Colombia, it fields a conjoint survey in which respondents evaluate hypothetical country-condition profiles that vary across political, economic, and social features. The survey measures individual-level thresholds, formalizes a threshold migration framework in the context of the Roy model, and estimates the elasticity of return. The study is pre-registered, with data collection ongoing.

Social and Distributional Effects of Capital Grants for Small and Medium Enterprises on Employers and Employees: Evidence from Post-War Iraq

Thu, 01 Jan 2026 00:00:00 +0000

This project studies how economic recovery affects socially cohesive attitudes and how its effects extend through social and economic networks via indirect exposure. In societies emerging from conflict, the benefits of growth and development may be unevenly distributed, creating or exacerbating inequalities. We study this process through the lens of a high-value capital grant program for small- and medium-sized enterprises in post-conflict Iraq, where material competition and grievances are a source of social tensions. The program supports business growth and employment creation with large capital grants (averaging $16,000), and we measure effects on socially cohesive attitudes using surveys of business owners, their employees, and owners’ social ties and professional peers. A pre-existing employee-employer register allows us to measure effects on the social attitudes of workers and to detect changes in socially cohesive attitudes among workers expected to benefit and lose as a result of the turnover induced by the program.

Endogenous household reorganization and social program manipulation

Mon, 01 Jan 2024 00:00:00 +0000

We study the effects of unconditional cash transfers to Syrian refugees in Lebanon on household composition and child migration between households. Using a regression discontinuity design, we do not find evidence that children regularly move from non-beneficiary households to those receiving benefits. While a survey of households appears to support the hypothesis that they reorganize based on whether they receive assistance, other data sources concerning the same households suggest otherwise. Instead, we find evidence that non-beneficiary households likely overreport the number of household members in an effort to increase the assistance they receive. This misreporting is predominantly driven by households that have recently stopped receiving program assistance, and more than 85% of the effect of treatment on reported household size is explained by non-beneficiary households overreporting the number of girls aged five and below. These results underscore the importance of considering the incentives for endogenous household reorganization and misreporting in the context of aid policies.