Activity Number:
|
78
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Business & Economics Statistics Section*
|
Abstract - #301298 |
Title:
|
Bayesian Analysis of Bank Closure Decisions
|
Author(s):
|
Viridiana Lourdes*+ and Omar Aguilar and Guillermo Rosas
|
Affiliation(s):
|
ITAM and Merrill Lynch Quantitative Research and ITAM
|
Address:
|
Rio Hondo #1 Col Tizapan Progreso, Mexico City, , 01000, Mexico
|
Keywords:
|
Duration models ; Survival analysis ; Bayesian analysis ; Bank crisis ; Dynamic models
|
Abstract:
|
A full Bayesian analysis is developed in this paper to explore the potential set of factors behind a bank closure decision. A general class of dynamic duration models for bank failures is developed and estimated, treating bank closings as an event timed by bank regulators and using time-varying correlated covariates over time-to-closure as explanatory variables. We review empirical findings in applying these models in a study of more than 100 banks in Mexico and Argentina during the past ten years. We discuss model assessment, residual analysis, and MCMC algorithms developed to fit these new class of models and conclude with comments on future potential developments together with model extensions. This is joint work with Omar Aguilar and Guillermo Rosas.
|