<abstract xmlns="http://www.w3.org/1999/xhtml"><p> Very high costs of systemic banking crises emphasize the importance of early warning models for these crises. In order to create an early warning model for systemic banking crises a combined approach is implemented. The first approach applied in this paper is signal approach, however, with some modifications as compared with its standard application in the literature. On the basis of individual indicators two composite indices are created. Unlike other papers in this field, the author has chosen a 24-month period before the beginning of the crisis as a signal horizon, while the signal horizon in the literature is usually considered to be a period of 12 months before and 12 months after the crisis onset. The second approach represents logit model whereas the independent variables are actually the indicators with the best performances obtained within the signal approach. In order to check the robustness of indicators, the Bayesian model averaging technique is used. The indicator that represents the credit growth rate, besides being a part of the composite index, is statistically significant in all estimated specifications of the logit model, including the technique of Bayesian model averaging. Additionally, trends in the international market have a significant influence on the domestic banking system and its stability, and hence also on the probability of occurrence of a systemic banking crisis.</p></abstract>

Very high costs of systemic banking crises emphasize the importance of early warning models for these crises. In order to create an early warning model for systemic banking crises a combined approach is implemented. The first approach applied in this paper is signal approach, however, with some modifications as compared with its standard application in the literature. On the basis of individual indicators two composite indices are created. Unlike other papers in this field, the author has chosen a 24-month period before the beginning of the crisis as a signal horizon, while the signal horizon in the literature is usually considered to be a period of 12 months before and 12 months after the crisis onset. The second approach represents logit model whereas the independent variables are actually the indicators with the best performances obtained within the signal approach. In order to check the robustness of indicators, the Bayesian model averaging technique is used. The indicator that represents the credit growth rate, besides being a part of the composite index, is statistically significant in all estimated specifications of the logit model, including the technique of Bayesian model averaging. Additionally, trends in the international market have a significant influence on the domestic banking system and its stability, and hence also on the probability of occurrence of a systemic banking crisis.

Predicting Systemic Banking Crises Using Early Warning Models: The Case of Montenegro

Journal of Central Banking Theory and Practice

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Very high costs of systemic banking crises emphasize the importance of early warning models for these crises. In order to create an early warning model for...

Predicting Systemic Banking Crises Using Early Warning Models: The Case of Montenegro

Online veröffentlicht: 23. Sept. 2017

Seitenbereich: 157 - 182

Eingereicht: 02. März 2017

Akzeptiert: 30. Mai 2017

DOI: https://doi.org/10.1515/jcbtp-2017-0025

Schlüsselwörter
early warning models, systemic banking crises, Montenegrin banking system, signal approach, composite index, logit model, Bayesian model averaging, synthesis of signal and logit approach, credit expansion, economic cycle

© Central Bank of Montenegro

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Predicting Systemic Banking Crises Using Early Warning Models: The Case of Montenegro

Online veröffentlicht: 23. Sept. 2017

Seitenbereich: 157 - 182

Eingereicht: 02. März 2017

Akzeptiert: 30. Mai 2017

DOI: https://doi.org/10.1515/jcbtp-2017-0025

Schlüsselwörterearly warning models, systemic banking crises, Montenegrin banking system, signal approach, composite index, logit model, Bayesian model averaging, synthesis of signal and logit approach, credit expansion, economic cycle

© Central Bank of Montenegro

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.

Schlüsselwörter
early warning models, systemic banking crises, Montenegrin banking system, signal approach, composite index, logit model, Bayesian model averaging, synthesis of signal and logit approach, credit expansion, economic cycle