Compositional Growth Models

TL;DR

This paper reviews compositional growth models that explain the statistical properties of growth in various systems by decomposing entities into independent units and unifying their behavior through a Gaussian mixture framework.

Contribution

It provides a comprehensive review and a unifying theoretical framework for understanding compositional growth models across different domains.

Findings

Models explain growth statistics of firms and GDP

Growth rate distributions can be described as mixtures of Gaussians

Unified framework connects diverse compositional growth phenomena

Abstract

We review models of compositional growth, which were introduced to explain the growth statistics of various quantities ranging from firm sizes to GDP. In these models, entities are decomposed into units that grow independently. Thus, the growth rate of the entity is the addition of the growth rates of the composing units, with possibly heterogeneous weights. We review such models and show that they can be understood through a unifying theoretical framework, explaining the resulting growth rate distributions using mixtures of Gaussians.