Store brands and retail grocery competition

• Date: 01 October 2018
• Author: Rong Luo
• DOI: http://doi.org/10.1111/jems.12254
• Published date: 01 October 2018

653

J Econ Manage Strat. 2018;27:653–668. wileyonlinelibrary.com/journal/jems © 2018 Wiley Periodicals, Inc.

Received: 21 September 2016 Revised: 2 December 2017 Accepted: 30 April 2018

DOI: 10.1111/jems.12254

ORIGINAL ARTICLE

Store brands and retail grocery competition

Rong Luo

Department of Economics, Univer-

sity of Georgia, Athens, GA,USA

(Email: rluo@uga.edu)

Abstract

Grocery retailers rank store brands as the most important factor that differentiates

them from their competitors. Retailer competition should have a stronger impact on

the more substitutable national brands than the more differentiated store brands. How-

ever, the literature has not studied the impacts empirically. In this paper, I quantify the

different impacts of retailer competition on national brands and store brands using the

scanner data of a U.S. chain retailer. I estimate a structural demand and supply model

that incorporates the differentiation effect and retailer competition. The results show

that national brand consumers are more likely to switch stores than store brands con-

sumers. By analyzing two counterfactual cases, I find that 1) if the retailer did not sell

store brands, its profit would decrease, and the loss would be greater in markets with

more competitors; and 2) if the retailer competition had increased, then the national

brands' retail prices would decrease more than the store brands' prices.

1INTRODUCTION

Grocery retailer store brands are imperfect cheap substitutes to national brands. They haveseen g reat success in manycountr ies.

According to Private Label Manufacturers Association (PLMA) data, total sales of store brands in U.S. supermarkets were

$62.5 billion in 2015 with a unit share at 22.9%, and store brands' sales market share in Europe is 38.3%, with 51.8% in the UK,

34.1% in France, and more than 40% in Germany.1The growth of store brands has drawneconomists' attention to study retailers'

strategic incentives to sell store brands.

However, one important and understudied retailer incentive is to create differentiation from competing retailers. As reported

in the 2015 Annual State of the Store Brands Industry, grocery companies rank “to serve as a differentiator that drives traffic

to stores and builds customer loyalty” as the most important role of store brands.2Unlike a national brand that is often sold by

different retailers, a store brand is exclusively sold by its retailer. Thus, as the retailer competition increases, a retailer will drop

the prices of the national brands more than those of the store brands. The retailer differentiation becomes more important as a

retailer faces more competitors. In this paper, I study these impacts using the scanner data of a U.S. retailer chain, Dominick's.

I focus on the ready-to-eat cereal category because it is the largest breakfast industry and the competition in this industry is

intense, with hundreds of national brands and store brands.3

To incorporate brand differentiation and retailer competition, I set up a random coefficient discrete-choice demand model

with two important modifications. First, I assume that a consumer's utility of the outside option depends on the number of

competing grocery retailers, which is different from the standard zero mean assumption in the literature. Second, I allow the

random coefficients for consumers' brand tastes to be correlated. This relaxes the independence restrictions in the literature. The

diagonal elements of the variance–covariance matrix of the random coefficients represent consumer loyalty to products, and

I am very grateful to Mark Roberts, Paul Grieco, Peter Newberry, and Joris Pinkse for their indispensable guidance. I also wouldlike to thank Yue Liu, Thomas

Quan, Pasquale Schiraldi, John Turner, participants at 2016 International Industrial Organization conference, and colleagues at UGA forhelpful suggestions.

This paper also benefited greatly from the insightful comments of the anonymous co-editor and referee. All errors are my own responsibility.

2JOURNAL OF ECONOMICS & MANAGEMENTSTRATEGY

654

the off-diagonal covariances represent the substitutability of products. The benefit of estimating the full matrix is to identify

the substitution pattern across products. In the supply model, Dominick's chooses the profit-maximizing zone-level prices for

all brands. Dominick's groups its stores into different zones.4Within a zone, each product has the same price across stores.

I estimate the demand model using generalized method of moments (GMM). Endogeneity issue arises in the price, and I use

instrumental variables to identify the model. With the demand model estimates, I compute the unit costs and estimate the supply

model using OLS. I then study two counterfactual cases. The first shows that Dominick's would reduce its national brands'

prices if it did not sell store brands and the national brands' wholesale prices were fixed. This implies that selling store brands

creates brand differentiation and increases national brand prices. I also find that store brands become more important as retailer

competition increases. In the second case, I assume that each Dominick's store faces 10% more competitors and solvefor the new

profit maximizing prices. The results show that Dominick's would drop both national and store brands' prices, but the decreases

in the national brands would be larger.

This paper contributes to the literature by first quantifying the different impacts of retailer competition on the national brands

and store brands. A few papers have studied store brands and retailers' competition theoretically. Liao and Yano (2013) analyze

a theoretical model to study how retailer competition affects the retail prices of store brands and national brands differently.

They are also interested in retailers' assortment decisions of store brands and national brands and how they change with retailer

competition. Corstjens and Lal (2000) also emphasize the store differentiation effect of store brands and find that store brands

contribute to higher profits. Other works that have focusedon retailer competition include Smith (2004), Matsa (2011), Dhar and

Hoch (1997), and Steenkamp and Geyskens (2013), who study the reasons for store brand penetration variation across retailers

and regions.

This paper also contributes to the literature on the incentives of retailers introducing store brands by emphasizing the cross-

retailer differentiation effect of store brands. The empirical literature has studied three other types of incentives of retailers

to sell store brands. First, many researchers have focused on whether the introduction of store brands increases a retailer's

bargaining poweragainst national brand manufacturers, including Mills (1995), Chintagunta, Bonfrer, and Song (2002), Pauwels

and Srinivasan (2004), Steiner (2004), Meza and Sudhir (2010), Cohen and Cotterill (2011), and Bontems, Monier-Dilhan, and

Réquillart (1999).

The second retailer incentive is to create product differentiation within its product set and increase the national brand markups.

Studies on this incentive include Pauwels and Srinivasan (2004), Ailawadi and Harlam (2004), Soberman and Parker (2004),

Gabrielsen and Sørgard (2007), and Cotterill and Putsis (2000). A few other related works (Dhar, Hoch, & Kumar, 2001; Du,

Lee, & Staelin, 2005; Hoch & Lodish, 1998; Soberman & Parker, 2006) analyze the impact of product differentiation, store

brands quality, and the pricing strategy on the profit of a retail store.

The third incentive is to increase consumers' store loyalty by introducing good quality store brands. The findings in the

literature (Ailawadi, Pauwels, & Steenkamp, 2008; Baltas & Argouslidis, 2007; Bonfrer & Chintagunta, 2004; Martos-Partal

& González-Benito, 2011; Seenivasan, Sudhir, & Talukdar, 2012) are mixed. Several papers were also interested in retailers'

strategic positioning decisions of store brands (Morton & Zettelmeyer, 2004; Raju, Sethuraman, & Dhar, 1995; Sayman, Hoch,

& Raju, 2002; Sethuraman, 2009).

The rest of the paper proceeds as follows. In Section 2, I describe the theoretical background of the paper.Section 3 describes

the data and some motivating facts about store brands and national brands. The demand model and supply model are explained

in Section 4. Section 5 discusses the identification of parameters and the estimation method. The estimation results are shown

in Section 6. In Section 7, I analyze two counterfactual cases. Section 8concludes the paper.

2THEORETICAL BACKGROUND ON THE IMPACT OF RETAILER

COMPETITION ON STORE BRANDS AND NATIONAL BRANDS

Liao and Yano (2013) study the impact of retailer competition on the prices of the national brands and store brands theoretically.

They consider a model of two geographically differentiated retailers, each selling an identical national brand and a differentiated

store brand. The quality of the two store brands can be different, and both are lower than the national brand. Consumers are

heterogenous in their location and their willingness to pay for quality. They prove that, compared with being a monopolist, 1) a

retailer lowers the prices of both the national brand and the store brand when it faces a competing retailer; and 2) the national

brand's price drops more than the store brand's.

The intuition of the two results is as follows. First, the retailer's prices of both products decrease because the demand for its

products becomes more elastic when competition increases. Thus, the retailer lower the prices to maximize profit. Second, the

two store brands are differentiated in both location and quality while the two national brands are differentiated only in location,