| This paper provides systematic analysis of the empirical literature about the relation between financial liberalization and banking crises by conducting a meta-analysis. To our knowledge, this is the first study based on a meta-analysis as a tool to identify banking crises’ origins, especially financial liberalization effect. We are interested in explaining heterogeneity of results reported in previous empirical works by investigating the importance of specific characteristics of studies, data, methodology and accounting for model uncertainty. Our contribution resides in the use of the Bayesian Model Averaging approach to determine potential explanatory factors of earlier finding discrepancies. We find that the sample size plays an important role in explaining divergences in previous results, and that the use of multidimensional measures in developing countries tends to reduce significantly the link between financial liberalization and banking crises. Furthermore, we find that some estimation techniques deliver results systematically different from those obtained via other methods. For example, the use of logit approach results in substantially important estimates of the relationship. |
Abstract
|
