Publications
- Ahumada, G., Otero, G., & Cabib, I. (2026). Neighborhood cohesion and depressive symptoms: Longitudinal evidence from Chile. Cities, 173. DOI: https://doi.org/10.1016/j.cities.2026.106899.
- Yáñez, R., Olcay, C., Navea, J., Ahumada, G. & Jara, B. (2025). Food insecurity through the lens of multidimensional poverty: Evidence from Chile. Social Science & Medicine, 383. DOI: https://doi.org/10.1016/j.socscimed.2025.118495.
- Ahumada, G., Cantillan, R. & Jara, B. (2024). Social capital and individual well-being in the post-disaster period: The case of Hurricane Maria in Puerto Rico. International Journal of Disaster Risk Reduction, 103. DOI: https://doi.org/10.1016/j.ijdrr.2024.104308.
- Ahumada, G., & Iturra, V. (2021). If the air was cleaner, would we be happier? An economic assessment of the effects of air pollution on individual subjective well-being in Chile. Journal of Cleaner Production. 289, 1-9. DOI: https://doi.org/10.1016/j.jclepro.2020.125152.
- Ahumada, G., Iturra, V. & Sarrias, M. (2020). We Do Not Have the Same Tastes! Evaluating Individual Heterogeneity in the Preferences for Amenities. Journal of Happiness Studies. 21, 53-74. DOI: https://doi.org/10.1007/s10902-019-00081-2. Stata codes and R scripts for replication: Download.
- Perez, D., & Ahumada, G. (2015). Cambio estructural y desindustrialización: evidencia para Colombia. Revista Venezolama de Gerencia. 20(71), 517-533. Available here.
- Ahumada, G., & Penso, L. (2014). Caracterización socioeconómica de la subregión del Canal del Dique. Revista Aguaita. 26, 1-22. Avilable here.
- Ahumada, G. & Martelo, E. (2010). La tenencia de la tierra en Colombia desde la conquista hasta 1936. Revista de Economía OIKOS. 24, 1-18. Available here.
Research Papers
Personal Networks and Mental Health: Longitudinal Evidence on the Role of Network Heterogeneity and Status Composition in Depressive Symptoms (with Ignacio Cabib and Gabriel Otero) - Submitted;
Research on personal networks and mental health has primarily focused on relationship frequency and type, overlooking two key structural dimensions: network heterogeneity and socioeconomic composition. This study analyzes how these dimensions shape depressive symptoms over time in Chile, a context marked by high depression rates and strong social stratification. Using data from the Longitudinal Social Study of Chile, we estimate dynamic panel models to address time-invariant confounding and reverse causality. Network attributes are measured using a position generator that captures occupational diversity and the presence of high- and low-status alters. Results suggest that network heterogeneity exerts a modest protective effect, but only with a temporal lag, indicating that structural diversity requires time to generate psychosocial benefits. Socioeconomic composition emerges as a stronger predictor: contact with high-status individuals reduces depressive symptoms, whereas greater exposure to low-status alters increases them. These effects vary by socioeconomic position, indicating the dual role of networks.
Research on personal networks and mental health has primarily focused on relationship frequency and type, overlooking two key structural dimensions: network heterogeneity and socioeconomic composition. This study analyzes how these dimensions shape depressive symptoms over time in Chile, a context marked by high depression rates and strong social stratification. Using data from the Longitudinal Social Study of Chile, we estimate dynamic panel models to address time-invariant confounding and reverse causality. Network attributes are measured using a position generator that captures occupational diversity and the presence of high- and low-status alters. Results suggest that network heterogeneity exerts a modest protective effect, but only with a temporal lag, indicating that structural diversity requires time to generate psychosocial benefits. Socioeconomic composition emerges as a stronger predictor: contact with high-status individuals reduces depressive symptoms, whereas greater exposure to low-status alters increases them. These effects vary by socioeconomic position, indicating the dual role of networks.
The Mental Health Returns to Compulsory Schooling: The case of Chile (with Ignacio Cabib and Gabriel Otero) - Submitted;
Education is a central social determinant of health, yet the causal effects of educational attainment on mental health remain contested and appear to vary across institutional contexts. This study examines the long‑run mental health returns to education in Chile, a highly unequal middle‑income country where this relationship has received limited causal attention. We use data from the 2023 National Survey on Health, Sexuality, and Gender and exploit Chile’s 2003 compulsory schooling reform—which extended mandatory education from eight to twelve years—as a source of exogenous variation in educational attainment. Age‑specific exposure to the reform is used as an instrumental variable for years of education. The reform increased educational attainment by 0.43 years, with strong first‑stage diagnostics. Two‑stage least squares estimates indicate that one additional year of education reduces the PHQ‑4 total score by 0.79 points (12% of the mean) and the depression subscale by 0.48 points (14% of the mean). The estimated effect on anxiety is negative but not statistically significant. Heterogeneity analyses indicate variation by gender and region, and exploratory analyses provide suggestive evidence consistent with pathways involving reproductive health knowledge, body satisfaction, and employment. These findings contribute to social epidemiological evidence on education as an upstream determinant of mental health in unequal middle‑income contexts.
Education is a central social determinant of health, yet the causal effects of educational attainment on mental health remain contested and appear to vary across institutional contexts. This study examines the long‑run mental health returns to education in Chile, a highly unequal middle‑income country where this relationship has received limited causal attention. We use data from the 2023 National Survey on Health, Sexuality, and Gender and exploit Chile’s 2003 compulsory schooling reform—which extended mandatory education from eight to twelve years—as a source of exogenous variation in educational attainment. Age‑specific exposure to the reform is used as an instrumental variable for years of education. The reform increased educational attainment by 0.43 years, with strong first‑stage diagnostics. Two‑stage least squares estimates indicate that one additional year of education reduces the PHQ‑4 total score by 0.79 points (12% of the mean) and the depression subscale by 0.48 points (14% of the mean). The estimated effect on anxiety is negative but not statistically significant. Heterogeneity analyses indicate variation by gender and region, and exploratory analyses provide suggestive evidence consistent with pathways involving reproductive health knowledge, body satisfaction, and employment. These findings contribute to social epidemiological evidence on education as an upstream determinant of mental health in unequal middle‑income contexts.
Blasting dust: The pollution-health Impact of industrial mining developments (with Nathaly Rivera, Nicolás Gómez-Parra, and Lenin H. Balza);
We study the pollution-health channel of the large-scale industrial mining developments leveraging exogenous variation in the geographical expansion of mineral leases in Chile over time and the distance to these installations. Using a battery of estimations in double differences and novel measures of expansion, we find a large negative impact of these developments on satellite-derived criteria air pollution concentrations within 50 km of these facilities, particularly SO2. Later on, we link this increased pollution with the health outcomes of nearby communities. These results have important implications for the debate on the local welfare impact of mining developments.
We study the pollution-health channel of the large-scale industrial mining developments leveraging exogenous variation in the geographical expansion of mineral leases in Chile over time and the distance to these installations. Using a battery of estimations in double differences and novel measures of expansion, we find a large negative impact of these developments on satellite-derived criteria air pollution concentrations within 50 km of these facilities, particularly SO2. Later on, we link this increased pollution with the health outcomes of nearby communities. These results have important implications for the debate on the local welfare impact of mining developments.
Work in Progress
Why Associational Life Fails to Build Generalized Trust under Inequality (with Roberto Cantillan and Vicente Espinoza);
Why does participation in voluntary associations fail to generate generalized trust consistently in unequal societies? We argue that the answer lies not in participation alone, but in the configuration and durability of memberships across social domains. Cross-domain portfolios are most conducive to generalized trust, yet they are also the hardest to sustain. By contrast, memberships clustered within a single domain can support localized trust without extending it to strangers. We show that generalized trust is highest among individuals with cross-domain memberships, lower among the isolated, and lowest among those in clustered portfolios. For neighborhood trust, however, clustered participation performs much better, approaching the levels associated with cross-domain memberships. We further show that cross-domain portfolios are the least stable over time, and that their persistence is socially stratified in ways consistent with structural precarity rather than a generic resource advantage. Under inequality, the civil society–trust gap reflects not a lack of participation, but the difficulty of sustaining the forms of participation most likely to generate generalized trust.
Why does participation in voluntary associations fail to generate generalized trust consistently in unequal societies? We argue that the answer lies not in participation alone, but in the configuration and durability of memberships across social domains. Cross-domain portfolios are most conducive to generalized trust, yet they are also the hardest to sustain. By contrast, memberships clustered within a single domain can support localized trust without extending it to strangers. We show that generalized trust is highest among individuals with cross-domain memberships, lower among the isolated, and lowest among those in clustered portfolios. For neighborhood trust, however, clustered participation performs much better, approaching the levels associated with cross-domain memberships. We further show that cross-domain portfolios are the least stable over time, and that their persistence is socially stratified in ways consistent with structural precarity rather than a generic resource advantage. Under inequality, the civil society–trust gap reflects not a lack of participation, but the difficulty of sustaining the forms of participation most likely to generate generalized trust.
