Profitability, capital, and risk in US commercial and savings banks: Re-examination of estimation methods

This study compares the relationships between the three main bank variables i.e., profitability, capital, and risk of US commercial and savings banks for the period 1995Q1-2015Q4. As the literature analyzes the relationship between these variables in pairs, thus suffering from an 'omitted variable bias' and a 'simultaneous equations bias', we examine whether these biases are statistically significant. We compare the common methodology of a three two-equation system with both a three-equation system and three separate OLS regressions. While there are significant differences in the coefficients of the three main variables within the sample period, we could not find substantial differences in out-of-sample forecast estimates.