Financial structure: the contribution of game theory and causality models.

Summary In addition to the practical and institutional interest, there are theoretical issues in conducting research on financial structure. Indeed, the diversity of models has led to a multiplicity of determinants. This has not allowed the enigma of capital structure to be resolved. Introduced by Donaldson [1961] and developed by Myers and Majluf [1984], hierarchical financing, or pecking order theory, provides new explanations and fills in some of the shortcomings of optimum models. It then appears as a general theory of financing. Subsequently, the use of game theory reveals that this was not the case when the beliefs and strategies of investors or company managers were taken into account. The work carried out in this thesis emphasizes the limitations of the hierarchical model, which is only partially valid in static terms. Moreover, its operation in dynamics is conditioned by the anticipation of a low information asymmetry and of future projects whose expected return is greater than that of current projects. Moreover, the existence of interdependent choices and of determinants that are not directly observable has led to a system of structural equations with latent variables. The objective of this approach is to translate economic concepts and to simultaneously capture the different debt rates. Empirical analyses conducted on 390 French industrial and commercial groups over the period 1987-1994 led to several results. First, a causal link appears from the short-term financial structure to the long-term ratio. Second, an adjustment behavior seems to emerge from short-term choices, but the results remain mixed at this level. Third, an atypical hierarchy is detected in the absence of information asymmetry. In the end, supplier credit functions as a substitute for bank debt.