A Bayesian nonparametric approach to modeling market share dynamics

TL;DR

This paper introduces a Bayesian nonparametric model for dynamic market share analysis, capturing firm interactions, entry/exit barriers, and technological costs, with a simulation method for economic regime transitions.

Contribution

It develops a novel Bayesian nonparametric framework using hierarchical mixtures and a generalized Pólya urn scheme for modeling complex market share dynamics.

Findings

Effective simulation of market regime transitions.

Model captures barriers to entry and firm interactions.

Infinite-dimensional properties relate to Fleming-Viot diffusions.

Abstract

We propose a flexible stochastic framework for modeling the market share dynamics over time in a multiple markets setting, where firms interact within and between markets. Firms undergo stochastic idiosyncratic shocks, which contract their shares, and compete to consolidate their position by acquiring new ones in both the market where they operate and in new markets. The model parameters can meaningfully account for phenomena such as barriers to entry and exit, fixed and sunk costs, costs of expanding to new sectors with different technologies and competitive advantage among firms. The construction is obtained in a Bayesian framework by means of a collection of nonparametric hierarchical mixtures, which induce the dependence between markets and provide a generalization of the Blackwell-MacQueen P\'{o}lya urn scheme, which in turn is used to generate a partially exchangeable dynamical particle system. A Markov Chain Monte Carlo algorithm is provided for simulating trajectories of the system, by means of which we perform a simulation study for transitions to different economic regimes. Moreover, it is shown that the infinite-dimensional properties of the system, when appropriately transformed and rescaled, are those of a collection of interacting Fleming-Viot diffusions.