ABSTRACT
With comparable coefficients and slightly higher R-squared values, Table 9 shows that the negative relationship between asset growth and changes in the core deposits-to-liabilities ratio indeed holds equally well for the sample including the crisis years. Also these outcomes are robust to estimation using quarterly data (yielding comparable coefficients and significance levels). Compared to Tables 6-8, explanatory power is much lower in Table 9. This is because the fixed effects in the panel estimation increase the explanatory power more in the levels specification (Tables 6-8) than in the specification using growth rates (Table 9).
