2. Bank Soundness: A PLS-SEM Approach

During the Global Financial Crisis (GFC) of 2007–2009, even banks in industrial economies with long established markets suffered significantly. This highlights, weaknesses in the banking system and the importance of a sound banking sector. This paper applies Partial Least Squares Structural Equation Modeling (PLS-SEM) to explain the drivers of bank soundness in the G7 countries during the period 2003–2013. PLS-SEM models are able to handle latent variables and complex models, and thus, PLS-SEM is suitable for this study. In creating a parsimonious model, the study assembles 17 manifest variables of six constructs as the direct cause and eight constructs as the indirect cause of bank soundness. The structural equation model comprises of six latent exogenous constructs [Capital (C), Asset (A), Management (M), Earnings (E), Liquidity (L) and Sensitivity (S)] which explains the observed consequences of bank soundness in these countries. Results indicate that CAMELS constructs are able to explain 32.5% of the variation in banks’ soundness. The model’s predictive relevance (Q ² ) in regards to endogenous construct stands at a medium category of 0.315. The results imply that banks placed high importance on off-balance sheet and capital activities, and thus, taking on higher risk. Surprisingly, banks were also operating at low levels of capital and liquidity, resembling banks that failed during the Great Depression of the 1930s. The weakness in capital and liquidity measures shows the need for policy makers to have a better understanding of sound banking, before quantifying measures and creating policies that makes banks’ less prone to crises episodes and create convergence with soundness.