Business Cycle Fluctuations, Economic Growth & Granular Behavior: An Agent-Based Approach

Abstract

This work proposes a generalization of Delli Gatti et al.’s (2005) model in order to study the granular behavior observed empirically (Gabaix, 2011), namely a few very large firms are able to account for a very high fraction of macroeconomic fluctuations. A sample made of 40000 Spanish firms during the period ranging from 1995 to 2015 is employed in order to carry out the empirical analysis. It is found that the original model with constant-return-to-scale technology is able to generate a degree of heterogeneity closer to that observed empirically than that produced by the model with decreasing-return-to-scale technology. Nevertheless, the lower degree of heterogeneity produced by the latter makes it possible to reduce the market share of the largest company, which, in the original model, is far from what is observed.