A time-series approach to exploring aggregate optimal capital structure: cointegration analysis

Abstract

Although there are a substantive number of corporate capital structure models which postulate the existence of an optimal capital structure, only Miller's model (1977) explicitly addresses the issue of an optimal aggregate capital structure. To date, there is very little empirical evidence to substantiate or refute the existence of an aggregate optimal capital structure. Johansen's procedure for cointegration testing is employed to test Miller's hypothesis, along with other theories of optimal capital structure. The cointegration approach is advantageous in that it allows for the testing of long-run equilibrium relationships between series which are non-stationary. The Johansen's procedure utilizes tests which have well-defined limiting distributions unlike those employed in the Engle and Granger procedure. Additionally, this methodology allows us to explicitly define the number of cointegrating vectors and, therefore, the number of common stochastic trends. The results indicate that there is an equilibri...