Documentos CEDE
Accede a las publicaciones que reúnen trabajos de profesores/as e investigadores/as de la Facultad de Economía, basados en información del Centro de Datos CEDE. Presentan análisis económicos y resultados preliminares que aportan evidencia y abren discusiones académicas sobre temas relevantes para el país.
Documento CEDE 2025-14
JEL: E44, G21, L13
In this paper, we analyze the competition in the Colombian banking sector using banklevel monthly balance sheet information. We estimate the changes in measures of market power due to the exogenous introduction of a liquidity regulation. Our results suggest that introducing a net stable funding ratio increased the Lerner index in the short term, thus signaling a higher exercise of market power. We rationalize these changes in a simple theoretical model that allows us to analyze the tightening of liquidity requirements for banks and find that a higher Lerner index implies more market power in the loan market than in the deposit market.
14-05-2025
Documento CEDE 2025-13
JEL: I10, I12, I18, Q50, Q54
This paper analyzes the impact of narrow changes in temperature exposure during pregnancy on maternal morbidity in Colombia. Using administrative individual-level health service records and high-resolution daily weather data at the municipality level, I find that the rate of healthcare utilization during pregnancy and delivery increases monotonically with temperature. Services like hospitalizations increase by 1.69% with each additional day above 29◦C compared to the reference bin of [21, 23)◦C. In contrast, services requiring less time under direct observation or that are primarily outpatient (i.e., emergency room visits, consultations, and medical procedures) increase by 1%. In a two-month exposure window, each additional day in the highest temperature bin is associated with nearly a 2% increase in hospitalizations, which is double the impact observed for other health services. Women from low socioeconomic conditions and living in rural areas seem to be more affected by temperature shocks, especially in high-level morbidity outcomes, like hospitalizations.
13-04-2025
Documento CEDE 2025-12
JEL: K42, K14, H41, H75, K23
¿Puede la regulación de los mercados de cannabis reducir el crimen? Este artículo presenta una revisión sistemática y un meta-análisis de la literatura de inferencia causal que busca responder esta pregunta. A partir de una muestra inicial de 31 estudios, sintetizamos los hallazgos de 9 artículos que cumplían con criterios rigurosos de inclusión. Adicionalmente, proponemos un marco conceptual que vincula la regulación del cannabis con la reducción del crimen a través de distintos mecanismos: la contracción del mercado ilícito, la disminución de las rentas criminales, la reducción del uso de la violencia como mecanismo para hacer cumplir acuerdos ilegales, así como la redistribución de recursos policiales y el fortalecimiento de intervenciones en salud pública. Los resultados del meta-análisis indican que la regulación del cannabis, en particular para uso medicinal, se asocia con una reducción del crimen en general, con efectos más pronunciados para delitos violentos. También encontramos heterogeneidad según el estatus de publicación, lo que sugiere la posible presencia de sesgos de publicación. En conjunto, los resultados sugieren que reformar los marcos regulatorios en torno al cannabis podría contribuir a disminuir la violencia y debilitar las estructuras del crimen organizado.
12-04-2025
Documento CEDE 2025-11
JEL: D14, I22, G53, H55
This paper investigates the impact of behavioral frictions and information provision on retirement planning through the Dual Advisory program in Colombia. The program gradually became mandatory to prevent costly mistakes and offered personalized information on switching pension plans. Using a regression discontinuity design and administrative data, we estimate the causal effects of the program on switching behavior and contribution patterns. The findings indicate that the program reduced unfavorable switches by approximately 29.5%, due to a combination of deterring would-be switchers for whom the switch would have resulted in lower expected pension (22.7%) and the information received by those who attended the sessions (6.7%). However, the program also deterred favorable switches by 10.7%. There is no evidence of the program affecting contribution frequency post-intervention.
11-03-2025
Documento CEDE 2025-10
JEL: D13, J16, J22, H41
The study of the effect of integrative care policies—as those that target the needs of caregivers and multiple care recipients—on time allocation for both women and men, is an empirical exercise not widely covered in the existing Care Economy literature, specially in diverse developing settings. I focus on the impact of the District Care System implemented by Bogot´a’s local government in 2020 through the study of Manzanas del Cuidado (MCs), as they offer well-being and human capital formation services to unpaid caregivers. For the period between 2019 and 2021, I estimate the Intention to treat (ITT) effect on time allocation in paid and unpaid labor, as well as labor market outcomes for women and men using an Event Study approach. Results indicate that women do not change their patterns of unpaid and paid time use, and while men exhibit a reduction in time allocated to paid work, estimates do not remain robust across various specifications, and identification issues may arise, limiting a conclusive causal interpretation attributable to the program. Labor market participation do not change for both women and men. Mechanisms such as the rigidity of gender social norms, the frequency of service usage, and methodological constraints may help unravel the reasons behind the observed effects.
10-02-2025
Documento CEDE 2025-09
JEL: I18, K42, C45
This text compiles ten years of experiences in developing machine learning models for the Colombian public sector. The main applications focus on public health and citizen security. In the first case, it demonstrates how to significantly improve the risk adjustment formula of the Capitation Payment Unit (UPC) and how to build a decision model to prevent unnecessary hospitalizations as a result of cost-effective home medical visits. In the second case, it identifies the causal effect of police presence in a quadrant of Bogotá and its impact on crime. To this end, it is necessary to estimate the elasticities ofpolice presence with respect to crime, cross-elasticities (indirect effects on Other quadrants). With this information it is possible to do counterfactual analysis and tin particular to estimate the impact on crime of the optimal police reassignment in the city. Additionally, a methodology is described to uncover the true crime rate in Bogotá, and underreporting of these events is estimated. Finally, a brief reflection is provided on the challenges faced by policymakers implementing these solutions.
09-02-2025
