Determinants of banks fragility in the MENA region using a dynamic model

Measuring the probability of default of commercial banks is considered a central issue for regulators as well as for practitioners. From previous literature that measure bank insolvency risks, most studies focused on the banking sector of developed countries and less attention is paid towards developing countries. The aim of this study is to identify the determinants of bank fragility of commercial banks in the MENA region. The selected variables rely on the CAMEL approach that is required by supervisory authorities. A Z-score technique was conducted on 13 banking systems and a sample of 239 banks that operate in the MENA region during the period 1999-2017. A dynamic model with a lagged variable was used in the System Generalized Method of Moments (SYS-GMM) estimator. In additional, two tests were conducted to validate our model specifications and to select the optimal model: Fisher's test to detect the homogeneity of the sample and Hausman test to identify the exogeneity of this specific effect on the explanatory variables. Overall, this study has revealed that large banks in MENA region are less vulnerable to default risk. The lagged variable of stability revealed a positive relationship which confirms an evolution in the stability of banks operating in the MENA region and can be explained by the introduction of new prudential rules and regulations. Moreover, capitalization level is showed a negative correlation with insolvency which supports the regulatory hypothesis. However, the imperative finding is that liquidity failed to show a significant relationship with insolvency. Finally, our results have shown that economic growth can be considered as an essential factor in reducing the fragility of banks operating in the MENA region. Although our study has revealed, a negative relationship between liquidity and the probability of default of commercial banks in the MENA region, unexpectedly it failed to prove the significance of this relationship. Probably, the persistent high level of liquidity that MENA banks enjoy, especially Gulf banks, prevented the model from detecting the evolution of this liquidity.