The retail sector has become of great importance for what it represents for the economic performance of a country, its central role in relations with upstream and downstream markets, and for the impact that its anticompetitive practices can have on social welfare.
Multiproduct by nature, the retail sector is characterized, on the one hand, by a complex configuration stemming from the multiplicity of forms, formats, products, and pricing and advertising strategies and, on the other hand, by its concentration, with a few large supermarket chains competing with small downtown shops. Whereas the former supply a wider range of products along with a variety of additional services (such as gas stations, shopping malls, restaurants and entertainment areas for kids), the latter generally offer narrower product lines but constitute an alternative to consumers through either higher quality, specialized offerings in the supply of specific categories of products (such as wine shops, vegetable markets, butcheries, etc.) or because prices are considerably lower (such as hard-discount stores).
For all these reasons, this sector has been of central attention for policy makers and economic research. A wide variety of topics has been treated (such as retailer competition, vertical relations, consumer product and supermarket choice, switching, search and shopping costs, the effects of new product introduction, the economics of private labels, etc.) from both theoretical and empirical perspectives in the Marketing and Economics literature. Yet many interesting questions remain to be answered. That is why I decided to concentrate my research in this area.
One interesting and extensively studied topic is the existence of the so-called private labels or store own-brands. Private labels (hereafter PL) are retailers' own-branded products supplied exclusively in their stores and produced by a separate manufacturer. They are produced as quality-equivalent products respective to regular manufacturer brands usually distributed in the whole territory (also known as national brands, NB). However, consumers often perceive them as of lower quality, which may explain why they are offered at lower prices in average. Evidence shows that PLs are supplied at a 20% lower price in average relative to a quality-equivalent NB (Berges-Sennou et al., 2009).
In France, a particular way to promote PL demand common to the different retail chains seems to be loyalty programs. Consumers subscribing to the program, get a loyalty card (carte fidelité) which they have to show every time they pass by the checkouts. This will give them right to benefit from permanent rebates on some selected PL and some other special promotions.
My first paper is precisely motivated by this observation: profit-maximizing retailers giving additional rebates on their lower priced own brands. To shed light on this fact, I empirically examine the effects of loyalty programs on PL demand.
Loyalty programs (hereafter LP) are present in almost all the retailing markets. Most work the same way: a member who purchases today gets a reward to be used next time she returns to the store (or after she crosses some threshold). Previous researchers have provided several explanations as to why retailers offer such costly programs. They can be summarized in two ways: 1) consumer retention, as they are more likely to come back when there is a promised price reduction; and 2) as a way to exercise market power, in particular, LPs can be used as an explicit discriminatory device as customers must subscribe to the LP to enjoy the benefits. As previously mentioned, the boost of the demand for a specific product or category of products seems to be another motive.
The interesting feature is that loyalty rebates are lagged, i.e. the discounts announced today on some products are accumulated as euros or ‘miles’ in customer’s account and after a given time/money threshold is crossed, the acquired amount of money is given back to customers as a purchase coupon to be expended in any of the retailer's stores. Theory shows that even when rebates are lagged to a subsequent period, consumers perceive current prices as lower. Additionally, theory predicts that firms are able to raise prices and get higher overall profits when they introduce loyalty programs.
I estimate brand-level demand using discrete-choice methods taking into account household membership to loyalty programs. I use a three-dimensional panel of quantities and prices for up to 13 brands of plain yogurt, purchased from the 6 largest supermarket chains in up to 94 departments of France, weekly in 2006. I also observe some demographic characteristics including household membership to supermarket LPs. In addition to the well documented challenges faced when estimating demand, such as the endogeneity of prices and the dimensionality problems implied by the large number of brands, I also deal with the correlation between membership of the supermarket LP and unobserved supermarket attributes.
Results confirm that private labels are, on average, less valued relative to NB. However, I find that the marginal valuation of PL products increases with subscription to the supermarket LP, which confirms the belief that LPs serve as a way to boost store-brand demand. Moreover, when customers subscribe to separate LPs of competing retailers, the expected effects are weaker, i.e. the marginal valuation of PL products decreases with the number of subscriptions and customers are more sensitive to price changes.
Another explanation for the retailers’ efforts for making customers loyal is to avoid them doing multi-stop shopping and concentrate all their purchases with only one retailer instead. This is precisely the focus of my second (ongoing) research project (joint with Daniel Herrera, also a Ph.D. at TSE).
People with similar characteristics may have quite different shopping preferences. On the one hand, some customers like to concentrate their purchases with only one retailer and take advantage of being loyal (benefit from loyalty rewards, for instance), time saving and the gains in experience when patronizing a particular retailer. On the other hand, some others prefer visiting multiple separate suppliers motivated by reasons such as to get the best deals, the existence of differentiated product lines, or due to some unanticipated event (a dinner party with the need for particular ingredients they do not regularly purchase or because they run out of some staple, say milk, before expected). We then say that two consumers have heterogeneous shopping patterns when they visit a different number of retailers within the same shopping period. Therefore, a consumer who goes to only one retailer within, say, a week will be a one-stop shopper and a consumer visiting several separate suppliers within the same week will be a multi-stop shopper.
This heterogeneity in consumers' shopping patterns may be determined by several factors such as preferences, demographics (income, age, household size, location, etc.), information frictions (search costs), differentiated retailers, and the time availability for shopping activities (opportunity cost of time). Theory has introduced a concept that summarizes most of them: shopping costs.
They are defined as all the consumers' real or perceived costs of using additional suppliers (Klemperer, 1992). More precisely, according to Chen and Rey (2012a), shopping costs reflect “the opportunity cost of time spent in traffic, parking, selecting products, checking out, and so forth”. Otherwise stated, shopping cost is defined as the opportunity cost of time when shopping.
It is an empirical fact that similar consumers may have different shopping patterns. In France, for example, around 85% of households exhibited multi-stop shopping behaviour in grocery shopping in 2005. One can argue that this heterogeneity in shopping patterns could respond mainly to individual preferences, and then the answer to any question regarding such a heterogeneity would be simple: similar consumers do different things just because they have different tastes for shopping. However, shopping costs may account also for the consumer preference for shopping. In fact, a common feature in the literature is that heterogeneous shopping patterns will always exist as long as there is a mix of consumers with heterogeneous shopping costs and retailers with differentiated product lines or specializations.
Economic theory has shown that in the presence of shopping costs, some features that would otherwise be considered good from a social welfare perspective can have adverse effects. The introduction of a new product variety, firms competing head-to-head by producing close substitutes and banning below-cost pricing of competitive products are some examples.
In fact, according to Klemperer (1992), in a competitive setting with differentiated product lines, retailers may be tempted to undercut prices in order to make one-stop shoppers become multi-stop ones by patronizing different retailers. On the other hand, if competition is head-to-head, by selling products as homogeneous as possible, customers will stay with only one retailer because the benefit of visiting an additional retailer will not compensate the shopping costs. As a consequence, competition is reduced and prices are higher.
Klemperer and Padilla (1997) show that in the presence of shopping costs the introduction of a new product can cause a decrease in rivals' profits generating the so-called “indirect business-stealing” effect. A firm offering more varieties than its rivals is attractive for consumers as they can get all the common and new products through a single retailer, rather than staying with the rival who only offers one product or doing multi-stop shopping and paying the extra cost of patronizing several retailers. As a consequence, the retailer introducing a new variety will sell more of all other products in his range making rivals' profits decrease. This leads to the introduction of too many varieties with respect to the socially optimal number.
Chen and Rey (2012a, b) show that retailers can exploit the fact that some consumers do like to be multi-stop shoppers to price discriminate. In this sense, they will never want to push competitors out of the market but keep them in instead, in order to be able to extract rents from multi-stop customers, by adopting loss-leading (when competing with smaller specialized rivals) or cross subsidization strategies (when competing against similar rivals).
All these striking theory findings and the lack of empirical literature on the topic motivated our interest in the structural identification of consumer shopping costs, with the primary objective of giving empirical support to the widespread assumption that differences of shopping costs explain the heterogeneity in consumer shopping patterns. And, second, to have an empirical tool that allow us to test some theory predictions and policy conclusions.
For the future I hope to continue to develop a research agenda on retailing and consumer shopping behaviour. In particular, I hope to work on topics such as information frictions, price advertising and consumer search costs.
 According to the European Commission (2010), it represents 4.2% of the European Union GDP, has around 17.4 million workers among other indicators. Also, food expenditure represents around 13% on average for European households.
 Some programs, such as Casino's, work slightly different as they give points (``miles'') to customers according to a predetermined exchange rate, and have a catalogue where members can pick a gift according to the cumulated number of miles.
 Source: TNS Worldpanel by TNS-Sofres (Kantar).
Berges-Sennou, Fabian, Bontems, Phillip and Requillart, Vincent (2009),“L’impact économique du développement des marques de distributeurs”, Working paper, Toulouse School of Economics, 30 p.
Chen, Zhijun and Rey, Patrick (2012a), “Loss-leading as an exploitative practice”, The American Economic Review, Vol. 102, No. 7, pp. 3462-3482.
________ (2012b), “Competitive Cross-Subsidization”, unpublished working paper.
European Comission (2010), Retail market monitoring report, http://ec.europa.eu/internal_market/retail/docs/monitoring_report_en.pdf.
Klemperer, Paul (1992), “Equilibrium product lines: competing head-to-head may be less competitive”, The American Economic Review, Vol. 82, No. 4, pp. 740-755.
Klemperer, Paul and Padilla, Jorge (1997), “Do firms’ product lines include too many varieties?” RAND Journal of Economics, Vol. 28, No. 3, pp. 472- 488.