Out-of-sample Predictability of Economic Efficiency Measures of U.S. Banks: Evidence of capital adequacy requirements

Abstract This paper contributes to the sparse debate on the effect of capital adequacy requirements on banks' economic efficiency measures. Precisely, we evaluate the out-of-sample predictability of capital adequacy requirements on banks' economic efficiency measures using Support Vector Regression (SVR) model with Linear, Polynomial and Radial Basis Function kernels and ordinary least squares (OLS) model. This analysis is important because a prediction of economic efficiency measures allows for an untangle view of bank's progress that is useful for management as it gains a high degree of transparency in the evaluation of future events. Our framework adapts optimization of h-block cross-validation to account for serial correlation of economic variables to produce robust sets of tuning parameters for SVR model. Using a total of 10,380 December quarterly observations of U.S Commercial and Domestic banks spanning from 2008 through 2019, empirical results show that SVR model provides better benchmarking in-sights in the evaluation of economic efficiency measures compared to the OLS model. Furthermore, in contrast to previous approaches identifying a single "best" model among competing models, the results of Model Confidence Test suggests that the out-of-sample forecasting confidently identifies superior predictive accuracy of SVR model-based forecasts over OLS model.