The Capital Ratio and the Interest Rate Spread: The Panel Threshold Regression Approach

Abstract

Using balanced annual observations of insured US commercial banks, this paper investigates the nonlinear impacts of the lagged capital ratio on the interest rate spread by employing the panel threshold regression model with one and two threshold variables which divide our sample into two and four regimes, respectively. The threshold variables we use are the change in capital ratio, the main components of change in the capital ratio, namely change in the capital or change in risk-weighted asset and ROE changes. In single threshold models, regimes that correspond to the higher change in capital ratio, higher contribution of capital to the change in capital ratio, and higher contribution of the risk-weighted asset to the change in the capital ratio show a stronger impact of the lagged capital ratio on the interest rate spread. In the case of using lagged ROE changes as threshold variable, surprisingly banks that are less exposed to the fall in ROE (as a result of increasing the capital ratio) tend more to raise the interest rate spread to offset the fall. In the panel threshold regression model with two threshold variables and four regimes, we find similar results.