Estimating Demand With Substitution and Intraline Price Spillovers

Abstract

Standard choice models used to optimize product line prices are effective at capturing the effects of prices on the affordability of products but do not incorporate intraline price spillovers, or the effects of prices on consumer preferences for branded product lines. To date, no approach exists to measure such spillovers, and thus their existence and characteristics are subject to controversy. Different authors have proposed that different product line items, here referred to as known value items (KVIs), generate price spillovers. Yet product line prices and quality levels are typically highly correlated, and thus it is not clear which spillovers should be modeled. No empirical approach exists to test for multiple price spillovers simultaneously while controlling for quality spillovers. This study proposes an empirical approach that includes a choice model to abstract rich demand patterns and an estimation algorithm that addresses the endogeneity and collinearity of KVI prices. An empirical application based on vehicle choice data illustrates how the approach can be implemented in practice and demonstrates the importance of modeling spillovers while yielding interesting managerial implications. The results establish that different KVIs generate separate spillovers and illustrate their distinct roles. KVI price spillovers can be economically important, inducing large cross-price elasticities that range from – 1.31 to .72. Furthermore, neglecting spillovers can bias own-price elasticity estimates by up to 15% and cross-price elasticities by more than 100%.The empirical results generate KVI-specific insights that managers in the automobile industry can use to improve their pricing strategies. For other industries, the manuscript proposes an approach and an implementation algorithm that managers can use to estimate spillovers and optimize product line prices. The empirical application illustrates the implementation of the approach and shows how managers can model price spillovers to increase profits by almost 3%.