Explaining differences in corporate capital structure: Theory and new evidence

Abstract

This paper tests a theory of optimal capital structure based on three simple imperfections: (i) personal and corporate taxes; (ii) bankruptcy related costs; and (iii) transfer of wealth from unprotected creditors. In contrast to other empirical tests (most of which rely on comparative statics to devise a linear regression), we derive a testable equation explicitly from the firm's value-maximizing first order condition. Our cross-sectional tests confirm earlier findings by Warner (1977) and Ang et al. (1982) that the fixed (explicit) costs of bankruptcy are not statistically significant.However, we find strong explanatory power in the marginal (implicit) costs of bankruptcy. These costs have two components. First, the data suggest that firms with high fixed-asset intensity carry heavier debt loads because of the floor that these assets create for potential bankruptcy costs to be borne by security-holders. This result explains inter-industry differences in debt-ratios found by Schwartz and Aronson (1967) and Scott (1972). It also explains some anomalies in recent tests by Bradley et al. (1984) who find a positive relationship between debt ratios and depreciation when their model predicted a negative relation. In our view, depreciation acts more asa proxy for fixed asset and less as a proxy for non-debt tax shield which Bradley et al. were analyzing. Secondly, our tests suggest that an unprotected creditor or wealth transfereffect provides additional stimulus to issuance of debt and that this stimulus is stronger than the net tax effect. Schwartz (1981) finds such distributional explanations to be both economically meaningful and worthy of further empirical research because of its potential for initiating legal reform. We modestly claim that our finding in this regard is an important first step in this research and is similar to the wealth-transfer effect from bondholders to stockholders studied by Masulis (1980, 1983).