Andrés Ham

Research

“Stress-testing survey-to-survey imputation: Understanding when poverty predictions can fail“, with Paul Corral, P.F. Lanjouw, Leonardo Lucchetti, and Henry Stemmler. World Bank Policy Research Working Paper WPS11192. August 2025. (R&R at World Bank Economic Review)

Accurate and timely poverty measurement is central to development policy, yet the availability of up-to-date high-quality household survey data remains limited—particularly in countries where poverty is most concentrated. Survey-to-survey(S2S)imputation has emerged as a practical response to this challenge, allowing practitioners to update poverty estimates using recent surveys that lack direct welfare measures by borrowing information from other comprehensive surveys. A critical review of the method is provided, revisiting its statistical underpinnings and testing its limitations through extensive model-based simulations. Through model-based simulations, the analysis demonstrates how violations of parameter stability, omitted variable bias, and shifts in survey design can introduce substantial errors—particularly when imputing across time or under economic and structural change. Results show that standard corrections such as re-weighting or covariate standardization may fail to eliminate these biases, especially when imputing across time or under structural change. The performance of alternative model specifications is also evaluated under various methods, including performance under heteroskedastic errors, non-normality. The findings offer practical guidance for practitioners on when S2S imputation is likely to succeed, when it should be reconsidered, and how to communicate its limitations transparently in the context of poverty monitoring and policy design.

“Impact Evaluation of an English as a Foreign Language Program in the Context of a Workforce Training Initiative in Colombia“, with S. García, S. Majerowicz, and D. Maldonado. (R&R at Education Evaluation and Policy Analysis) [Working paper version in Spanish]

This study examines the effectiveness of an English language learning program for low-income working-age individuals in Bogotá, Colombia. We take advantage of an unexpected change in the admission protocol for an English language program offered within a vocational training initiative by the city government to estimate the causal effects of the program using two empirical strategies: a matching estimator and a regression discontinuity design. We find that being offered a spot in the program increased English scores by 0.2 standard deviations (SD), while for those students who attended at least 20% of the classes, the effect was 0.27 SD. The program was effective regardless of student gender or socioeconomic status, with no significant differences by individual characteristics. However, students whose teachers had above-average lesson delivery quality had significantly higher language gains (0.35 SD). The magnitude of the effects is substantial compared to the average results of education programs, equivalent to an improvement of between one-fifth and one-third of a level in the Common European Framework of Reference for Languages. Our results contribute to the limited evidence on effective interventions that improve English proficiency among adults, particularly those from vulnerable contexts.

“Is drug-related violence fueling emigration from Central America?“, with Leonardo Bonilla-Mejía, Jessica Bracco, and Leonardo Peñaloza-Pacheco. IZA Discussion Paper No. 18028. July 2025. (Submitted)

We study how drug-related violence affects emigration from Central America, a region with rapidly rising migration to the United States. Using multiple data sources, we apply an instrumental variables strategy based on proximity to drug-trafficking routes and coca production in Colombia. We find that violence significantly increases intentions, plans, and preparations to emigrate—especially to the U.S.—with stronger effects among young and high-skilled individuals. Mediation analysis suggests this response is driven by declining economic activity and, more importantly, deteriorating labor market conditions caused by escalating violence.

“Design without delivery: analysing and rethinking social protection in Honduras“. WIDER Working Paper 2025/101.

We analyse the disconnect between the intended design and actual outcomes of Honduras’s social protection system. The current framework results in low coverage rates and limited redistribution, and imposes an implicit tax on formal activity. It also discourages productivity-enhancing decisions by workers and firms, contributing to the persistence of high informality and inequality. To address these issues and fulfil the objectives of social protection, we propose reforms which aim to expand coverage, improve equity, and strengthen incentives for workers and firms to comply with norms. As Honduras discusses social protection reform, this study offers timely insights to inform policy.

“The labor market effects of drug-related violence in a transit country“, with Juanita Ruiz and Andrea Velásquez. IZA Discussion Paper No. 17126. July 2024. (New draft in preparation )

We estimate the effects of drug-related violence on individual labor market outcomes in a transit country. Transit countries do not have enough market power to determine the global supply or demand of drugs yet must deal with the consequences from drug trafficking activities. We implement a Bartik-type instrumental variables strategy which assumes that violence in Honduran municipalities located along drug transport routes changes when coca production in Colombia grows or contracts. Our results show that drug-related violence has negative effects on extensive and intensive margin labor market outcomes for transit country workers and has greater effects on women than men.

“Can’t Stop the One-Armed Bandits: The Effects of Access to Gambling on Crime” with Nicolas Bottan and Ignacio Sarmiento-Barbieri. (New draft in preparation)

We study the effect of a large increase in access to gambling on crime by exploiting the expansion of video gambling terminals in Illinois since 2012. Even though video gambling was legalized by the State of Illinois, local municipalities were left with the decision whether to allow it within their jurisdiction. The City of Chicago does not allow video gambling, while many adjacent jurisdictions do. We take advantage of this setting along with detailed incident level data on crime for Chicago to examine the effect of access to gambling on crime. We use a difference-in-differences strategy that compares crime in areas that are closer to video gambling establishments with those that are further away along with the timing of video gambling adoption. We find that (i) access to gambling increases violent and property crimes; (ii) these are new crimes rather than displaced incidents; and (iii) the effects seem to be persistent in time.

Resting papers