Abstract
The influence of firm maturation on firms' tax-induced financing and investment decisions
Empirical evidence supporting the predicted positive relation between firms' financial leverage and their marginal tax rates has been elusive. Scholes and Wolfson (1989a) argue that refinancing costs that accumulate with age affect the time-series variation in firms' tax-induced financing and investment policies. They predict that as corporate capital structures gradually become more constrained over time, firms' financing decisions will become less sensitive to changes in their tax rates. Secondly, Scholes and Wolfson predict that as firms are increasingly impeded from adjusting their capital structures, they will resort to relying more in investment-related tax shields.
This thesis tests Scholes and Wolfson's (1989a) predictions using panel data spanning firms' first through ninth years since their initial public offerings. Strong, robust evidence on the evolution in firms' debt and asset tax shields is consistent with their predictions. The empirical testing in this thesis begins with the estimation of the time-series pattern in firms' rate of reversion to their optimal capital structures, which suggests that adjusting leverage becomes more difficult over time. The longitudinal evidence also indicates that the positive relation between leverage and marginal tax rates subsides as firms mature. Further, firms are observed to gradually shift more toward investment-related tax shields when re-financing costs increasingly constrain their capital structures.
This thesis contributes to the literature by developing a research design that represents a richer empirical characterization of the capital structure problem by isolating the influence of dynamic adjustment costs on tax-induced financing and investment decisions. In addition, this thesis provides the first evidence on the time-series properties of firms' responses to stable tax shield incentives. Extant capital structure research examines either cross-sectional variation or time-series variation relating to changes in certain macroeconomic conditions such as tax laws.
Finally, this thesis answers Slemrod and Shobe's (1990) and Shevlin's (1999) call for empirical research that more precisely models non-tax costs and the time-series behavior of firms' reactions to tax incentives, respectively. This is largely achieved by exploiting the features of panel data to control for unobserved firm-specific effects to avoid omitted variable bias and to refine the estimation of within-firm dynamics.