Vertical Channel Analysis of the U.S. Milk Market

Abstract

The research in this paper extends the standard random coefficient logit demand (S-RCL) framework using a Box-Cox transformation of the attribute space (BC-RCL) to empirically investigate vertical conduct scenarios in milk manufacturing and retailing. S-RCL relies on ad hoc linear indirect utility that assumes constant marginal utility of attributes. The BC-RCL model relaxes this assumption allowing the data to determine the functional form of utility. Because using product-level data does not allow for an analytical solution for the power transformation parameters, we propose an algorithm to estimate the transformation and consumer heterogeneous taste parameters altogether. The supply selection bias associated with linear indirect utility is shown to have formidable policy implications. Compared to S-RCL, retailers under BC-RCL demand are shown to have more market power using a different best-fitted supply scenario. Elasticity and super-elasticity estimates are presented.