International monetary instability: an approach in terms of agents' interactions and information heterogeneity.

Summary This work aims to show the limits of linear models derived from the asset market approach - most often used in the study of exchange rate determination - and to contribute to the proposal of alternative models in an essentially heuristic approach. To do so, we start from two essential questions: the first concerns the opposition between linear determinist models and stochastic models and the second is the micro-macro transition. The first question, dealt with in chapter i, is reflected on the one hand by the study of two representative models of the asset market approach (the Dornbusch model and the portfolio model), and on the other hand by the study of stochastic processes used in financial analysis (the arch type models and Brownian motion). The first chapter concludes, on the one hand, by asserting the insufficiency of the stochastic approach in terms of non-stationarity, and on the other hand, by the impossibility of studying financial exchange rate series by linear determinist models. This leads to the question of the possibility of explaining exchange rate instability by non-linear determinist models, which calls into question the traditional assumptions about the behavior of agents, in other words, the assumptions of rational expectations and market efficiency. The idea implicit in these assumptions is the homogeneity of agents' expectations, while the opposite idea, that of information heterogeneity, implies non-linearity and constitutes a problem that falls under the micro-macro transition, discussed in chapter ii. This chapter discusses the theoretical and empirical factors conditioning the validity of the two hypotheses in question, as well as the criticisms from the point of view of the bounded rationality and mimetic approaches. Since these approaches fail to account for the problem of information heterogeneity, since they use tools derived from classical determinism, which is the very basis of the hypotheses we wish to challenge, we turn, in the third chapter, to the proposal of an alternative form of analysis - that of chaos theory - by making its connection with the analysis of complex dynamic systems. The criticism made in the first two chapters of this paper is not sufficient to justify this proposal.