Competition in Multi-Characteristics Spaces: Hotelling Was Almost Right

Working Paper: CEPR ID: DP1446

Authors: Andreas Lrmen; Jacques-François Thisse

Abstract: Lancasterian models of product differentiation typically assume a one-dimensional characteristics space. We show that standard results on prices and locations no longer hold when firms compete in a multi-characteristics space. In the location game with n characteristics, firms choose to maximize differentiation in the dominant characteristic and to minimize differentiation in the others when the salience coefficient of the former is sufficiently large. Thus, the principle of minimum differentiation holds for all but one characteristic. Furthermore, prices do not necessarily fall when products get closer in the characteristics space because price competition is relaxed when products are differentiated enough in the dominant characteristic.

Keywords: product positioning; multicharacteristics space; agglomeration

JEL Codes: L1; M3; R3


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
Maximizing differentiation in the dominant characteristic (L15)Relaxed price competition (L11)
Minimizing differentiation in other characteristics (L15)Less intense price competition (D49)
Closer products in the characteristics space (C52)Equilibrium prices do not necessarily fall (D59)

Back to index