Evaluación del riesgo operativo en empresas con distribución de alimentos y bebidas mediante un modelo de riesgo colectivo discreto
| dc.contributor.advisor | Giraldo Gómez, Norman Diego | |
| dc.contributor.author | Martinez Bolivar, Mateo | |
| dc.contributor.orcid | Norman Giraldo-Gómez [0000000152405358] | |
| dc.date.accessioned | 2026-01-19T16:03:51Z | |
| dc.date.available | 2026-01-19T16:03:51Z | |
| dc.date.issued | 2025 | |
| dc.description.abstract | Este estudio desarrolla y aplica un marco metodológico basado en el modelo de riesgo colectivo en tiempo discreto para la evaluación cuantitativa del riesgo operativo (RO) en empresas del sector real. El enfoque se centra en modelar la frecuencia y severidad de los eventos de riesgo, permitiendo proyectar la distribución de pérdidas agregadas y estimar reservas financieras para cubrir pérdidas esperadas, mediante la combinación de métodos actuariales clásicos y modelos estadísticos avanzados capaces de capturar sobredispersión, asimetría y dependencia temporal. Partiendo del Modelo de Riesgo Colectivo (CRM) en tiempo discreto, se implementa el enfoque Poisson-Gamma Separado (SPGA) para modelar de forma independiente la frecuencia de eventos y la severidad de las pérdidas. Este marco se amplía mediante la incorporación de Modelos Lineales Generalizados (GLM) y sus extensiones mixtas (GLMM), con el fin de integrar covariables operativas y controlar la heterogeneidad no observada. Para superar las limitaciones de los modelos estáticos ante la presencia de dependencia temporal, se exploran extensiones dinámicas: i) modelos G-ARMA para capturar autocorrelación en series de conteo y severidad, ii) el modelo Tweedie, que integra frecuencia y severidad en una sola estructura paramétrica con masa en cero, y iii) Modelos Generalizados de Scores Aditivos (GAS), que permiten la evolución adaptativa de los parámetros en el tiempo. La metodología se aplica a datos de siniestralidad vehicular (conjunto ausautoBI8999) debido a la confidencialidad de los datos reales del sector de alimentos y bebidas. Los modelos se validan mediante diagnósticos de ajuste, pronósticos y la estimación de medidas de riesgo basadas en la teoría de la ruina: probabilidad de insolvencia en horizontes finitos ($\psi(u,n)$), el Valor en Riesgo y el valor de cola en riesgo del supremo de pérdidas acumuladas ($VaR_\kappa$, $TVaR_\kappa$), la medida de parte negativa esperada (ENP) y el índice de Lundberg--Aumann--Serrano (LAS). Los resultados evidencian que los modelos dinámicos, en particular SPGA-GAS y Tweedie, ofrecen pronósticos estables y realistas, constituyéndose en una base sólida para la gestión del capital económico. Finalmente, se propone una hoja de ruta para el diseño e implementación de un Sistema de Administración de Riesgo Operativo (SARO) alineado con la normativa colombiana, integrando los modelos desarrollados en un sistema de monitoreo, alerta temprana y asignación de reservas. Se concluye que el enfoque de riesgo colectivo en tiempo discreto, enriquecido con modelos estadísticos flexibles, constituye una herramienta robusta y generalizable para la gestión proactiva del riesgo operativo en sectores no financieros expuestos a eventos de baja frecuencia y alto impacto. (Texto tomado de la fuente) | spa |
| dc.description.abstract | This study develops and applies a methodological framework based on the discrete-time Collective Risk Model (CRM) for the quantitative assessment of operational risk (OR) in firms from the real sector. The approach focuses on modeling the frequency and severity of risk events, allowing for the projection of aggregate loss distributions and the estimation of financial reserves to cover expected losses, by combining classical actuarial methods with advanced statistical models capable of capturing overdispersion, asymmetry, and temporal dependence. Building on the discrete-time Collective Risk Model, the Separated Poisson–Gamma Approach (SPGA) is implemented to independently model event frequency and loss severity. This framework is extended through the incorporation of Generalized Linear Models (GLM) and their mixed extensions (GLMM), in order to integrate operational covariates and control for unobserved heterogeneity. To overcome the limitations of static models in the presence of temporal dependence, several dynamic extensions are explored: (i) G-ARMA models to capture autocorrelation in count and severity series; (ii) the Tweedie model, which integrates frequency and severity within a single parametric structure with a point mass at zero; and (iii) Generalized Autoregressive Score (GAS) models, which allow parameters to evolve adaptively over time. The methodology is applied to vehicle insurance claims data (the ausautoBI8999 dataset) due to the confidentiality of real data from the food and beverage distribution sector. Model validation is conducted through goodness-of-fit diagnostics, forecasting performance, and the estimation of ruin-based risk measures, including the probability of insolvency over finite horizons ($\psi(u,n)$), Value at Risk and Tail Value at Risk of the supremum of accumulated losses ($VaR_\kappa$, $TVaR_\kappa$), the Expected Negative Part (ENP), and the Lundberg–Aumann–Serrano (LAS) index. The results show that dynamic models—particularly SPGA-GAS and Tweedie—provide stable and realistic forecasts, offering a solid quantitative basis for economic capital management. Finally, the study proposes a roadmap for the design and implementation of an Operational Risk Management System (ORMS) aligned with Colombian regulatory standards, integrating the developed models into a framework for monitoring, early warning, and reserve allocation. It is concluded that the discrete-time collective risk approach, enriched with flexible statistical models, constitutes a robust and generalizable tool for proactive operational risk management in non-financial sectors exposed to low-frequency, high-impact events. | eng |
| dc.description.curriculararea | Estadística.Sede Medellín | |
| dc.description.degreelevel | Maestría | |
| dc.description.degreename | Magíster en Ciencias - Estadística | |
| dc.format.extent | 1 recurso en línea (89 páginas) | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.instname | Universidad Nacional de Colombia | spa |
| dc.identifier.reponame | Repositorio Institucional Universidad Nacional de Colombia | spa |
| dc.identifier.repourl | https://repositorio.unal.edu.co/ | spa |
| dc.identifier.uri | https://repositorio.unal.edu.co/handle/unal/89245 | |
| dc.language.iso | spa | |
| dc.publisher | Universidad Nacional de Colombia | |
| dc.publisher.branch | Universidad Nacional de Colombia - Sede Medellín | |
| dc.publisher.faculty | Facultad de Ciencias | |
| dc.publisher.place | Medellín, Colombia | |
| dc.publisher.program | Medellín - Ciencias - Maestría en Ciencias - Estadística | |
| dc.relation.references | Aven, T. (2015). Risk analysis. Wiley: https://doi.org/10.1002/9781119057819. | |
| dc.relation.references | Barbiero, A., y Hitaj, A. (2024). Discrete half-logistic distributions with applications in reliability and risk analysis. Annals of Operations Research, 340, 27-57. | |
| dc.relation.references | Benjamin, K., y Konstantinos, F. (2002). Regression models for count time series. John Wiley & Sons, Ltd. doi: https://doi.org/10.1002/0471266981. | |
| dc.relation.references | Benjamin, M. A., Rigby, R. A., y Stasinopoulos, D. M. (2003). Generalized autoregressive moving average models. Journal of the American Statistical Association, 98(461), 214–223. Descargado 2025-07-24, de http://www.jstor.org/stable/30045208 doi: https://doi.org/ 10.1198/016214503388619238 | |
| dc.relation.references | Cameron, A., y Trivedi, P. (2013). Regression analysis of count data. Cambridge University Press. Descargado de https://books.google.com.co/books?id=d6BTfLBbxjcC. | |
| dc.relation.references | Chen, T., Desmond, A. F., y Adamic, P. (2023). Generalized additive modelling of dependent frequency and severity distributions for aggregate claims. Journal of Statistical and Econometric Methods, 12(4). | |
| dc.relation.references | Cossette, H., Marceau, E., y Maume-Deschamps, V. (2010). Discrete-time risk models based on time series for count random variables. ASTIN Bulletin, 40(1), 123–150. doi: https://doi.org/ 10.2143/AST.40.1.2049221. | |
| dc.relation.references | Cossette, H., Marceau, E., Trufin, J., y Zuyderhoff, P. (2020). Ruin-basedriskmeasuresindiscretetimeriskmodels. Insurance: Mathematics and Economics, 93, 246–261. | |
| dc.relation.references | Cossette, H., Étienne Marceau, y Toureille, F. (2011). Risk models based on time series for count random variables. Insurance: Mathematics and Economics, 48(1), 19-28. doi: https:// doi.org/10.1016/j.insmatheco.2010.08.007. | |
| dc.relation.references | Creal, D., Koopman, S. J., y Lucas, A. (2008). A general framework for observation driven timevarying parameter models. Tinbergen Institute Discussion Paper No. 08-108/4. Descargado de http://dx.doi.org/10.2139/ssrn.1297183. | |
| dc.relation.references | Creal, D., Koopman, S. J., y Lucas, A. (2013). Generalized autoregressive score models with applications. Journal of Applied Econometrics, 28(5), 777–795. | |
| dc.relation.references | Czado, C., Kastenmeier, R., Brechmann, E. C., y Min, A. (2012). A mixed copula model for insurance claims and claim sizes. Scandinavian Actuarial Journal, 2012(4), 278–305. doi: https://doi.org/10.1080/03461238.2010.546147. | |
| dc.relation.references | De Melo, M. A. B., Fernandes, C. A. C., y de Melo, E. F. L. (2018). Forecasting aggregate claims using score-driven time series models. Statistica Neerlandica, 354b. doi: 10.1111/ stan.12139. | |
| dc.relation.references | Dickson, D. C. M., y Waters, H. R. (1992). The probability and severity of ruin in finite and infinite time. ASTIN Bulletin, 22(2), 177–190. doi: 10.2143/AST.22.2.2005114. | |
| dc.relation.references | Dobson, A. J. (2010). An introduction to generalized linear models, second edition. Taylor & Francis. Descargado de https://books.google.com.co/books?id=0CAgx5kQSwcC. | |
| dc.relation.references | Escalante, C. C., y Arango, G. O. (2004). Aspectos básicos del modelo de riesgo colectivo. Matemáticas: Enseñanza Universitaria, XII(2), 3-15. | |
| dc.relation.references | Franco Arbeláez, L. C., y Velásquez Ceballos, E. (2010). Alternativas fundamentales para cuantificar el riesgo operacional. Ecos de Economía, 14, 8-43. | |
| dc.relation.references | Frees, E., Derrig, R., y Meyers, G. (2014). Predictive modeling applications in actuarial science. Cambridge University Press. Descargado de https://books.google.com.co/ books?id=QHrsAwAAQBAJ. | |
| dc.relation.references | Frees, E. W., y Wang, P. (2006). Copula credibility for aggregate loss models. Insurance: Mathematics and Economics, 38(2), 360–373. Descargado de https://doi.org/10.1016/ j.insmatheco.2005.10.004. | |
| dc.relation.references | Garrido, J., Genest, C., y Schulz, J. (2016). Generalized linear models for dependent frequency and severity of insurance claims. Insurance: Mathematics and Economics, 70, 205-215. doi: https://doi.org/10.1016/j.insmatheco.2016.06.006. | |
| dc.relation.references | Jeong, H., Valdez, E. A., Ahn, J. Y., y Park, S. C. (2021, sep 29). Generalized linear mixed models for dependent compound risk models. Variance, 14(1). | |
| dc.relation.references | Jiménez Moscoso, J. A. (2022). Introducción a la teoría estadística del riesgo (Primera edición ed.). Universidad Nacional de Colombia. | |
| dc.relation.references | Jørgensen, B., y Paes-DeSouza, M. C. (1994). Fitting tweedie’s compound poisson model to insurance claims data. Scandinavian Actuarial Journal, 1994(1), 69–93. Descargado de https://doi.org/10.1080/03461238.1994.10413930 doi: 10.1080/03461238.1994 .10413930. | |
| dc.relation.references | Klugman, S., Panjer, H., y Willmot, G. (2012). Loss models: From data to decisions. Wiley. Descargado de https://books.google.com.co/books?id=z0qdRiK7I_gC. | |
| dc.relation.references | Lee, W., Park, S. C., y Ahn, J. Y. (2019). Investigating dependence between frequency and severity via simple generalized linear models. Journal of the Korean Statistical Society, 48, 13–-28. Descargado de https://doi.org/10.1016/j.jkss.2018.07.003. | |
| dc.relation.references | Lefèvre, C., y Loisel, S. (2008). On finite-time ruin probabilities for classical risk models. Scandinavian Actuarial Journal, 2008(1), 41-60. doi: 10.1080/03461230701766882. | |
| dc.relation.references | Lesław, G., y Rudź, M. (2013). Sharp approximations of ruin probabilities in the discrete time models. Scandinavian Actuarial Journal, 2013(5), 352-382. doi: 10.1080/03461238.2011 .618761. | |
| dc.relation.references | McNeil, A. J., Frey, R., y Embrechts, P. (2005). Quantitative risk management: Concepts, techniques and tools. Princeton University Press. | |
| dc.relation.references | McNeil, A. J., Frey, R., y Embrechts, P. (2015). Quantitative risk management: Concepts, techniques and tools. Princeton University Press. | |
| dc.relation.references | International Association of Insurance Supervisors, I. A. (2024). Insurance core principles (Inf. Téc.). IAIS. Descargado de https://www.iaisweb.org/page/supervisory-material/insurance-core -principles. | |
| dc.relation.references | Panjer, H. H. (2006). Operational risk: Modeling analytics. Wiley. | |
| dc.relation.references | Peter, K. D., y Gordon, K. S. (2018). Generalized linear models with examples in r. Springer New York, NY. doi: 10.1007/978-1-4419-0118-7. | |
| dc.relation.references | Peña, A., Bonet, I., Lochmuller, C., Patiño, H., Chiclana, F., y Góngora, M. (2018). A fuzzy credibility model to estimate the operational value at risk using internal and external data of risk events. Knowledge-Based Systems. doi: 10.1016/j.knosys.2018.06.007. | |
| dc.relation.references | Privault, N. (2022). Notes on stochastic finance (Inf. Téc.). Nanyang Technological University. | |
| dc.relation.references | Quijano Xacur, O. A., y Garrido, J. (2015). Generalised linear models for aggregate claims: to tweedie or not? European Actuarial Journal, 5, 181-202. doi: https://doi.org/10.1007/ s13385-015-0108-5. | |
| dc.relation.references | Reinhard, J. M., y Snoussi, M. (1997). On the severity of ruin in the discrete risk model. Technical Report 74. | |
| dc.relation.references | Reinhard, J. M., y Snoussi, M. (2002). The severity of ruin in a discrete semi-markov risk model. Stochastic Models, 18(1), 85-107. doi: 10.1081/STM-120002776. | |
| dc.relation.references | Rydberg, T. H., y Shephard, N. (1999). Modelling the arrival of events in the presence of volatility. Journal of Applied Probability, 36(1). | |
| dc.relation.references | Sandström, A. (2006). Solvency: Models, assessment and regulation. Chapmann & Hall. | |
| dc.relation.references | Segura-Gisbert, J., Lledó, J., y Pavía, J. M. (2025). Dataset of an actual motor vehicle insurance portfolio. European Actuarial Journal, 15, 241–253. doi: https://doi.org/10.1007/s13385 -024-00398-0. | |
| dc.relation.references | Shi, H., y Wang, D. (2014). An approximation model of the collective risk model with inar(1) claim process. Communications in Statistics - Theory and Methods, 43(24), 5305–5317. doi: 10.1080/03610926.2012.729636. | |
| dc.relation.references | Shi, P., Feng, X., y Ivantsova, A. (2015). Dependent frequency–severity modeling of insurance claims. Insurance: Mathematics and Economics, 64, 417-428. Descargado de https://www.sciencedirect.com/science/article/pii/S0167668715001183 doi: https://doi.org/10.1016/j.insmatheco.2015.07.006. | |
| dc.relation.references | Song, P. X. k. (2007). Correlated data analysis: Modeling, analytics, and applications (Primera Edición ed.). Springer New York, NY. doi: https://doi.org/10.1007/978-0-387-71393-9. | |
| dc.relation.references | Steutel, F. W., y van Harn, K. (1979). Discrete analogues of self-decomposability and stability. The Annals of Probability, 7(5), 893–899. Descargado 2025-09-10, de https://www.jstor .org/stable/2243313 doi: 10.1214/aop/1176994950. | |
| dc.relation.references | Superintendencia Financiera de Colombia. (2006). Sentencia c-048 & c-049 de 2006. (Disponible en: https://www.superfinanciera.gov.co/publicaciones/20053/normativanormativageneralboletin- minhacienda-capitulo-superintendencia-financieraboletin-minhaciendacapitulo- superintendencia-financiera-diciembre-20053/). | |
| dc.relation.references | Triantafyllopoulos, K. (2008). Dynamic generalized linear models for non-gaussian time series forecasting. arXiv:0802.0219 [https://arxiv.org/abs/0802.0219]. (Preprint). | |
| dc.relation.references | West, M., y Harrison, J. (1997). Bayesian forecasting and dynamic models (2nd ed.). New York: Springer. | |
| dc.relation.references | Wood, S. (2025). mgcv: Mixed gam computation vehicle with automatic smoothness estimation [Manual de software informático]. Descargado de https://cran.r-project.org/ package=mgcv (R package version 1.9-3). | |
| dc.relation.references | Wüthrich, M. V. (2006). Premium liability risks: modeling small claims. Mitteilungen, 2006(1), 27-38. | |
| dc.relation.references | Wüthrich, M. V., y Merz, M. (2023). Statistical foundations of actuarial learning and its applications (1st ed.). Springer Cham. doi: https://doi.org/10.1007/978-3-031-12409-9. | |
| dc.relation.references | Xiang, H., y Jing, Y. (2025). A combined integer-valued autoregressive process with actuarial applications. Journal of Computational and Applied Mathematics, 459. doi: https://doi.org/ 10.1016/j.cam.2024.116384. | |
| dc.rights.accessrights | info:eu-repo/semantics/openAccess | |
| dc.rights.license | Reconocimiento 4.0 Internacional | |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | |
| dc.subject.ddc | 310 - Colecciones de estadística general | |
| dc.subject.ddc | 330 - Economía::332 - Economía financiera | |
| dc.subject.lemb | Distribución de poisson | |
| dc.subject.lemb | Modelos lineales (Estadística) | |
| dc.subject.lemb | Probabilidades | |
| dc.subject.lemb | Industrias alimentarias | |
| dc.subject.proposal | Riesgo operativo | spa |
| dc.subject.proposal | Modelo de riesgo colectivo en tiempo discreto | spa |
| dc.subject.proposal | GLM | eng |
| dc.subject.proposal | GLMM | eng |
| dc.subject.proposal | G-ARMA | eng |
| dc.subject.proposal | SPGA | eng |
| dc.subject.proposal | Distribución tweedie | spa |
| dc.subject.proposal | GAS | eng |
| dc.subject.proposal | pérdidas agregadas | spa |
| dc.subject.proposal | Probabilidad de ruina | spa |
| dc.subject.proposal | SARO | spa |
| dc.subject.proposal | Operational risk | eng |
| dc.subject.proposal | Discrete-time collective risk model | eng |
| dc.subject.proposal | Tweedie distribution | eng |
| dc.subject.proposal | aggregate losses | eng |
| dc.subject.proposal | Ruin probability | eng |
| dc.subject.wikidata | Analisis de riesgo | |
| dc.title | Evaluación del riesgo operativo en empresas con distribución de alimentos y bebidas mediante un modelo de riesgo colectivo discreto | spa |
| dc.title.translated | Operational risk assessment in food and beverage distribution companies using a discrete collective risk model | eng |
| dc.type | Trabajo de grado - Maestría | |
| dc.type.coar | http://purl.org/coar/resource_type/c_bdcc | |
| dc.type.coarversion | http://purl.org/coar/version/c_ab4af688f83e57aa | |
| dc.type.content | Text | |
| dc.type.driver | info:eu-repo/semantics/masterThesis | |
| dc.type.redcol | http://purl.org/redcol/resource_type/TM | |
| dc.type.version | info:eu-repo/semantics/acceptedVersion | |
| dcterms.audience.professionaldevelopment | Administradores | |
| dcterms.audience.professionaldevelopment | Investigadores | |
| dcterms.audience.professionaldevelopment | Investigadores | |
| dcterms.audience.professionaldevelopment | Maestros | |
| dcterms.audience.professionaldevelopment | Público general | |
| oaire.accessrights | http://purl.org/coar/access_right/c_abf2 |
Archivos
Bloque original
1 - 1 de 1
Cargando...
- Nombre:
- Tesis de Maestría en Ciencias - Estadística.pdf
- Tamaño:
- 1.21 MB
- Formato:
- Adobe Portable Document Format
Bloque de licencias
1 - 1 de 1
Cargando...
- Nombre:
- license.txt
- Tamaño:
- 5.74 KB
- Formato:
- Item-specific license agreed upon to submission
- Descripción:

