BAYESIAN IDENTIFICATION OF THE PREDICTORS FOR CAPITAL STRUCTURE IN EGYPT

Mohamed A. Ismail (U. A. E.) and Tarek I. Eldomiaty (U. A. E.)

Abstract

The practice of capital structure decisions is dominated by three theories of capital structure; trade-off, pecking order, and free cash flow. None of the three theories presents a complete answer to the determinants of capital structure. This study utilizes the Bayesian "Stochastic Search Variable Selection" (SSVS) procedure to identify the predictors of corporate capital structure. The SSVS procedure avoids the overwhelming burden of computing the posterior distribution over all possible models by employing the Gibbs sampler to simulate a sample from the posterior distribution and to identify the best subset, which is the one with highest frequency in the sample. The results show that the practice of capital structure decisions in Egypt is affected by the determinants of two theories; trade-off theory and pecking-order theory.

Keywords and phrases: capital structure, modelling, stochastic variable selection.