A Maxtent Model for Macroscenario Analysis
Advances in Complex Systems, Vol. 11, No. 5, pp. 719-744, 2008
Posted: 8 Dec 2009
Date Written: October 2008
In this paper, starting from Jaynes' MaxEnt methodology [10, 11], we follow the original idea of Aoki  to implement a canonical MaxEnt inference model for the replication of industrial firms' dynamics over a space of economic states. We develop an aggregate model to infer the distributions of agents at meso level using representative states. In particular, we estimate the access probability for agents in different states consistently with macroscopic economic constraints. The model is calibrated on the basis of a sample of firms, drawn from the AMADEUS database, within the manufacturing industry made up of nine sectors of economic activity from 1995 to 2004, and results come to experimental proof at aggregate macroscopic level.
Keywords: Statistical mechanics, canonical ensemble, MaxEnt, Gibbs distribution, econophysics, Cobb-Douglas technology
Suggested Citation: Suggested Citation