Diseño de un framework de clasificación supervisada para mejorar la gestión de cobranza de los asociados de la cartera microfinanzas de una cooperativa financiera
La cartera de microcréditos registra los mayores niveles de riesgo para las entidades financieras en comparación con otras unidades de negocio, son créditos para clientes con bajos ingresos, patrimonio limitado y no ofrecen garantías que respalden la operación contractual. Cuando estos incumplen o r...
- Autores:
-
Granda Rodriguez, Oscar Anibal
Niño Hernandez, Juan Manuel
- Tipo de recurso:
- http://purl.org/coar/version/c_b1a7d7d4d402bcce
- Fecha de publicación:
- 2016
- Institución:
- Universidad Industrial de Santander
- Repositorio:
- Repositorio UIS
- Idioma:
- spa
- OAI Identifier:
- oai:noesis.uis.edu.co:20.500.14071/35512
- Palabra clave:
- Microfinanzas
Riesgo De Crédito
Cobranza
Default De Cartera
Scoring De Seguimiento
Análisis Discriminante
Regresión Logística.
The microloan portfolio has the highest level of risk for financial institutions compared to other business units as they are credits for customers with low income
limited patrimony and don´t provide guarantees to support the contractual operation and
when they fail or they are late in the payments require greater use of collection tools. Microcredit clients pay their obligation by a few days late and still intensity in the collection is very high causing upset in the borrower affecting future business relationships and excesses of operating loads collection for the entity that generates little effectiveness of collection strategies and limited resource allocation. This work improves collection strategy in microfinance portfolio of a cooperative financial institution in nature by using statistical tools. It starts from a base of partners (customers) with sociodemographic historical information
financial variables
granting and credit behavior
from which it´s explained the probability that a customer in default. The entire process to determine the improvement in collection strategy using statistical methods generates the design of a framework covering ten steps. Initially it part from the selection of a target portfolio
in this case the business unit microfinance
it is defined the historical frame of information
the explanatory variables are obtained and the information is purged
then the default or failure is calculated as that there is no single criterion for defining which client is good and which client is bad; then the variables are analyzed with descriptive statistics. Then it uses statistical tools as Classification Trees
Discriminant Analysis and Logistic Regression was applied using SPSS software
the model that best explains the data using diagnostic test is selected. Subsequently
a collection scoring is designed by calculating the probability of default distributed in percentiles or "score distribution" that give an expected risk to finally design a differential collection strategy.
- Rights
- License
- Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)