Empirical evidence on determinants of capital structure

Based on received financial theory, we empirically examine the role of the following firm-specific determinants of leverage: bankruptcy costs, growth, variability, non-debt tax shields, collateral value, profitability, and size. For our sample, to focus on firm-specific aspects, we purposely use pooled time-series cross-sectional data from a single industry (electric and gas utilities) for the twenty-year period, 1975–1994. Our findings largely support theory, with the important exception of variability. We find that leverage is positively related with variability, contrary to the literature. We conjecture that this seemingly perverse relationship may be due to the yet unrecognized effect of variability: firms with greater variability of earnings have a greater chance that their non-debt tax shields may prove to be inadequate, and are therefore expected to take on higher levels of leverage.