What Threatens Tunisian Banking Stability? Bayesian Model Versus Panel Data Analysis

This paper aims to investigate the main determinants of Tunisian bank stability. To achieve this goal; we have used a dataset of ten (10) Tunisian banks during the period 1990-2015. These banks are the most dynamic and the most involved in the financing of the economy. The econometric strategy used in this paper was based on two approaches. The first one performed the Bayesian Model Average (BMA) to detect the most important indicators influencing bank stability. The second one was based on panel data analysis involving random effect regression. Results of these two methods have indicated that Tunisian bank stability is more sensitive to capital adequacy ratio, liquidity risk and the interaction between credit risk and liquidity risk. The capital adequacy ratio is positively and highly significantly associated with the dependent variable (Z-Score). However, liquidity risk and interaction variables exert a negative and significant effect on bank stability. These results have important policy implications. Banks and policy makers should continue to strengthen the capital adequacy ratio since it significantly contributes to improving bank stability. However, they should pay attention to liquidity risk as the main determinant of bank instability