Banks in trouble? A Early Warning System for the Prevention of Banking Crisis
Ana Fernández-Sainz / Felipe Llaugel
Universidad del País Vasco (UPV/EHU) / Universidad Autónoma de Santo Domingo
The regulatory and supervisory financial authorities have tried various methods to find an effective procedure in developing an early warning system of banking crises. Logistic regression models have been used but have shown some weaknesses, so we need new and better methods. The banking crisis occurred in the Dominican Republic between 2002 and 2004 has been used to compare the effectiveness of the logistic regression method over the use of Support Vector Machines (SVM) for the detection of banking crisis. In the analysis 30 financial indicators are used to determine which ones are most appropriate to build a model able to classify banks. In this context, the SVM method produced better results than logistic regression, in detecting problem banks and contradict the findings of other studies that ask about the ineffectiveness of the financial indicators to identify banking crises in emerging economies.
KeyWords: Banking crisis; logistic regression; support vector machines
DOI: 10.5295/cdg.100239af
Bibliographic: FERNÁNDEZ, A. y LLAUGEL, F.: "¿Bancos con Problemas? Un Sistema de Alerta Temprana para la Prevención de Crisis Bancarias", Cuadernos de Gestión, Vol 11, nº 2, pp.149-168, DOI: 10.5295/cdg.100239af
JEL Classification: C19, C44, C61

<< Back
  Cláusula de Extensión de Responsabilidad y Declaración sobre los Derechos de Autor
Diseño gráfico: FIDEM S.L. - Diseño técnico y programación: CDM Consultores