Testing Capital Structure Theories: Are the Models’ Assumptions Correctly Specified? Carmen Cotei Department of Economics, Finance and Insurance University of Hartford, 200 Bloomfield Ave., West Hartford, CT, 06117, USA E-mail: firstname.lastname@example.org Joseph Farhat Department of Finance Central Connecticut State University, 1615 Stanley St., New Britain, CT 06050, USA E-mail: email@example.com
In this study, we investigate the models used in testing the trade-off and pecking order theories. Specifically, we examine the symmetric behavior assumption and homogeneous coefficient assumptions. To test the adequacy of these assumptions we utilize a spline regression model. For the trade-off model, our results show that symmetrical rate of adjustment assumption is rejected across all industries. Firms tend to adjust faster toward the target leverage when they are above the target compared to when they are below the target. For the pecking order model, the test results reject the symmetric behavior assumption at the industry level as well as across all industries. Firms have the tendency to reduce debt by a significantly higher proportion when they have financing surplus compared to the proportion of debt issued when they have financing deficit.
Keywords: Capital structure; pecking order theory; trade-off theory
The growing literature of evaluating the efficiency of the trade-off theory versus pecking order theory has produced mixed evidence. Shyam-Sunder and Myers (1999) find more supportive evidence for the pecking order theory than for trade-off theory. Byoun and Rhim (2005) find that both theories explain significant variations in the firms’ total debt. Fama and French (2002) find evidence in favor and against both of the theories. Frank and Goyal (2003) find evidence inconsistent with the pecking order theory, especially for small firms. Lemmon and Zender (2004) find no supporting evidence for the trade-off theory, yet the costs of adverse selection were not able to explain the pecking order financing behavior that they documented. Flannery and Rangan (2006) find no supporting evidence for the pecking order theory. While most of the above studies use the same common empirical model to test both theories, none empirically examines the validity of the models’ assumptions. This study complements the extant literature on the pecking order theory and trade-off theory. Our focus is to investigate the adequacy of both specification and assumptions of the models used in testing the trade-off and pecking order theory. Previous empirical studies that use the partial adjustment model and the pecking order model developed by Shyam-Sunder and Myers’ (1999) have
European Journal of Economics, Finance And Administrative Sciences - Issue 11 (2008)
two implicit assumptions1. The first assumption is the symmetric behavior assumption. Under the partial adjustment model, firms adjust toward their optimum level of debt (target leverage) with the same rate regardless if they are above or below the target. This implies that the cost and benefit of being above the target leverage is identical to the one of being below the target. However, the trade-off theory does not predict such a behavior. The pecking order model predicts a symmetric behavior for firms with financing deficit (shortage of internal sources of funds) or financing surplus (excess of internal sources of funds). In other words, firms with financing deficit issue debt to finance their new investment, whereas firms with financing surplus end up retiring debt rather than repurchasing equity (Shyam-Sunder and Myers, 1999). The second assumption is the homogeneous coefficient assumption. The partial adjustment model employed to examine the trade-off theory assumes that firms across industries...