Too Close for Comfort? Understanding Peer Effects in Large Franchised Networks
31 Pages Posted: 27 Sep 2021
Date Written: August 26, 2021
Proximity between stores in franchise networks that share the same brand is a highly contentious issue. A high-performance store may have different impacts on its proximal peer stores from a negative impact, such as potential cannibalization, to a positive impact, such as improved brand reputation or knowledge spillover. Most of the previous studies exploring the peer effects between stores in branded franchise networks emphasize only either the positive effect or the negative effect. In this paper, we consider both and seek to understand the net peer effect between each store pair in a large franchise network and how it changes with distance. We adapt an approach from spatial econometrics utilizing a weight matrix and extend it to allow for identification of both positive and negative peer effects between stores within a large-scale network. By applying our method to a large dataset from a major national fast food restaurant chain in the US, we show that the peer effects are negative between stores within 3 miles of each other, positive between stores within 3-7 miles of each other, both of which are economically significant. However, the peer effects generally diminish beyond 7 miles. These results inform individual franchisees of when they may be threatened or benefit from their proximal peer stores that share the same brand and for lawmakers to design policies to protect individual franchisees from encroachment. These results also reconcile the seemingly conflicting view of the issue in the literature and in practice. We then conduct a network analysis based on our estimated peer effects. Our network analysis results provide guidance for the franchisor on which areas they should avoid for further expansion and which stores they should prioritize to mitigate the tension between the franchisor and the franchisees.
Keywords: Large-scale Franchise Network, Proximity Effect, Peer Effects, Semi-parametric Econometrics
Suggested Citation: Suggested Citation