Demand Models With Random Partitions

58 Pages Posted: 25 Jun 2018 Last revised: 21 Mar 2019

See all articles by Adam N. Smith

Adam N. Smith

University College London - UCL School of Management

Greg M. Allenby

Ohio State University (OSU) - Department of Marketing and Logistics

Date Written: March 12, 2019

Abstract

Many economic models of consumer demand require researchers to partition sets of products or attributes prior to the analysis. These models are common in applied problems when the product space is large or spans multiple categories. While the partition is traditionally fixed a priori, we let the partition be a model parameter and propose a Bayesian method for inference. The challenge is that demand systems are commonly multivariate models that are not conditionally conjugate with respect to partition indices, precluding the use of Gibbs sampling. We solve this problem by constructing a new location-scale partition distribution that can generate random-walk Metropolis-Hastings proposals and also serve as a prior. Our method is illustrated in the context of a store-level category demand model where we find that allowing for partition uncertainty is important for preserving model flexibility, improving demand forecasts, and learning about the structure of demand.

Keywords: Bayesian inference, location-scale family, Polya urn, Markov chain Monte Carlo, price elasticity

Suggested Citation

Smith, Adam N. and Allenby, Greg M., Demand Models With Random Partitions (March 12, 2019). Available at SSRN: https://ssrn.com/abstract=3192926 or http://dx.doi.org/10.2139/ssrn.3192926

Adam N. Smith (Contact Author)

University College London - UCL School of Management

One Canada Square
London, E14 5AA
United Kingdom

Greg M. Allenby

Ohio State University (OSU) - Department of Marketing and Logistics ( email )

Fisher Hall 524
2100 Neil Ave
Columbus, OH 43210
United States

Here is the Coronavirus
related research on SSRN

Paper statistics

Downloads
80
Abstract Views
590
rank
326,077
PlumX Metrics