Ex-post analysis of the evolution of financial indexes tends to match, in a causal relationship, ... more Ex-post analysis of the evolution of financial indexes tends to match, in a causal relationship, exogenous events contained in an information stream, and observed market reactions. It is obvious that market players may react to globally available information, the effects of a self-reinforced endogenous mechanism. This may destabilize the global system even in the absence of an external perturbation, and may play a major role in explaining some observed stylized facts. Indeed, empirical analysis of financial markets has shown a number of these stylized facts, such as heavy tails or volatility "bursts," which are difficult to explain in terms of the evolution of fundamental economic variables. Indeed, the non-Gaussian, non-stable character of empirical distributions, such as excess demand or stock returns, demonstrates the weakness of any "independent agent" approach to modeling the real market. Starting with existing literature on the characterization of the behavior of random economies with many interacting agents, we identify a set of micro-economic interaction rules, which could help to explain the macro-economic observed market behavior. Following the work of Bornholdt and extending that of Brock and Durlauf, we will consider interacting agents whose payoff exhibit both a strategic complementarity with their nearest-neighbors actions and an eventual global substitutability with the global market state. In this setup we reconstruct a price process related to the imbalance between buyers and sellers. Finally, we investigate how the frustration resulting from the tendency of local imitation, with an additional coupling with the average 114 Transparency, Governance and Markets state of the system, reproduces the main observed stylized facts of real financial markets. We show how in this framework even the largest crash may emerge as a natural intrinsic metastable dynamic of the system induced by a collective phenomena, such as crowd effects or "herd" behavior.
Ex-post analysis of the evolution of financial indexes tends to match, in a causal relationship, ... more Ex-post analysis of the evolution of financial indexes tends to match, in a causal relationship, exogenous events contained in an information stream, and observed market reactions. It is obvious that market players may react to globally available information, the effects of a self-reinforced endogenous mechanism. This may destabilize the global system even in the absence of an external perturbation, and may play a major role in explaining some observed stylized facts. Indeed, empirical analysis of financial markets has shown a number of these stylized facts, such as heavy tails or volatility "bursts," which are difficult to explain in terms of the evolution of fundamental economic variables. Indeed, the non-Gaussian, non-stable character of empirical distributions, such as excess demand or stock returns, demonstrates the weakness of any "independent agent" approach to modeling the real market. Starting with existing literature on the characterization of the behavior of random economies with many interacting agents, we identify a set of micro-economic interaction rules, which could help to explain the macro-economic observed market behavior. Following the work of Bornholdt and extending that of Brock and Durlauf, we will consider interacting agents whose payoff exhibit both a strategic complementarity with their nearest-neighbors actions and an eventual global substitutability with the global market state. In this setup we reconstruct a price process related to the imbalance between buyers and sellers. Finally, we investigate how the frustration resulting from the tendency of local imitation, with an additional coupling with the average 114 Transparency, Governance and Markets state of the system, reproduces the main observed stylized facts of real financial markets. We show how in this framework even the largest crash may emerge as a natural intrinsic metastable dynamic of the system induced by a collective phenomena, such as crowd effects or "herd" behavior.
Uploads
Papers by gira l'elica