Resumo

PEREZ GARCIA, Jorge Iván; LOPERA CASTANO, Mauricio  e  VASQUEZ BEDOYA, Fredy Alonso. Estimation of bankruptcy risk probability in Colombian companies from a model for rare events. Cuad. Adm. [online]. 2017, vol.30, n.54, pp.7-38. ISSN 0120-3592.  https://doi.org/10.11144/javeriana.cao30-54.eprqe..

In order to discriminate bankruptcy risk and no bankruptcy risk of Colombian companies that reported their financial statements to the Superintendency of Corporations for the time period 2011-2015, this paper considers bankruptcy as a rare event and it employs a logistic model, a generalized additive model, a generalized extreme value model, and a binary generalized extreme value additive model (BGEVA). In comparative terms, the BGEVA model presents better predictive performance compared to the other models by assuming an extreme value distribution in the link function and semi-parametric structures in its estimations, thus allowing to establish the existing relation between default probability and explanatory variables.

Palavras-chave : bankruptcy; rare events; prediction models.

        · resumo em Português | Espanhol     · texto em Espanhol     · Espanhol ( pdf )