واکنش قیمت فعلی به ورودی جدید :مورد سوپر مارکت های ژاپنی

Large-scale supermarkets have rapidly expanded in Japan over the past two decades, partly because of zoning deregulations for large-scale merchants. This study examines the effect of supermarket openings on the price of national-brand products sold at local incumbents, using scanner price data with a panel structure. Detailed geographic information on store location enables us to define treatment and control groups to control for unobserved heterogeneity and temporary demand shock. The analysis reveals that stores in the treatment group lowered their prices of curry paste, bottled tea, instant noodles, and toothpaste by 0.4–3.1% more than stores in a control group in response to a large-scale supermarket opening.

The retail sector has been regarded as one of Japan’s least productive industries. In 2000, the McKinsey Global Institute issued a very influential report, which found Japan’s overall retail productivity is half of the US’s; in particular, the productivity of small-scale retail stores is only 19% of that in the US. The report points out that the large share of unproductive small retail shops was the main cause of overall low productivity. The report claims that this lower productivity hurt Japanese consumers through high prices. Since the report’s issuance in 2000, the structure of Japanese retail industries has changed dramatically. Fig. 1 displays the recent changes of the share by medium- and large-scale food stores, as well as total sales in Japan.1 The figure clearly shows that medium- to large-scale food stores increased their presence in Japan. The numbers of large food stores and mom-and-pop shops are reported in Fig. 2, which indicates that since 1991, the number of small food stores has decreased by about 50%, while medium-large stores have increased by about 20%.Although there are various reasons for the changes, one of the most influential causes was the deregulation of store locations at the national level. Small retail shops in Japan had been protected from competition with large retail shops by governmental regulation. Under the Large-Scale Retail Store Law (Daikibo Kouri Tenpo Ho), which was enacted in 1974, potential supermarkets entrants with a floor area of 500 or more square meters had to obtain permission from local incumbent merchants, as well as confirmation from local authorities. That is, the entry of large retail shops that would compete with local stores was heavily regulated. In 2000, the Large-Scale Retail Store Location Law (Daikibo Kouri Tenpo Ricchi Ho) replaced the Large-Scale Retail Store Law. This new law dropped the requirement for the local merchant union’s agreement for approval, and local authorities almost automatically approved new stores if the applications could prove that the new stores would not harm the local community’s environment, for example, by causing excessive noise or traffic jams, through an environmental assessment report. In response to this deregulation, openings of new large retail stores increased dramatically. Whether this rapid expansion of large retail shops benefited consumers through lower prices remains an empirical question. Studies on the effect of large supermarket entry on local pricing are rapidly emerging. Basker (2005) examines the effect of Wal-Mart openings on the pricing of local incumbents, using a city-level, quarterly panel price survey from the US. She selected 10 national brand items and found that Wal-Mart openings reduced the city’s average price of several products by 1.5–3%. A follow-up study by Basker and Noel (2009), based on panel data, again reports a price reduction effect of 1–2%. Hausman and Leibtag, 2009a and Hausman and Leibtag, 2009b report that Wal-Mart sells identical food items 15–25% lower than traditional supermarkets. Lira et al. (2007) examine the effect of opening new supermarkets on the local price index of 15 food-related items and find that local prices are reduced by 7–11%, based on Chilean data. Manuszak and Moul (2008) examine the case of office supply stores in the US and report that a higher density of store locations in a local area results in lower prices, after controlling for the endogeneity of store locations. They identify the endogeneity of local store density because stores are located in areas with higher demand and correct for this endogeneity using distance from the supply chain headquarters as an instrumental variable. Matsuura and Motohashi (2005) document the establishment-level dynamics of entry and exit for the Japanese retail sector and report the exit of establishments with lower labor productivity and the entry of establishments with higher potential for growth in labor productivity between 1997 and 2002. Their study clearly suggests that deregulation was efficiency enhancing, but it does not address its effect on prices that could lead to consumer welfare improvement, because their data set does not contain detailed information on prices. Ariga et al. (2001) report very frequent price changes of curry pastes sold at a supermarket in Japan and propose an original model of dynamic price discrimination, in which the supermarket discriminates among its customers with heterogeneous reservation prices by varying selling prices over time. Their empirical findings suggest the strategic nature of pricing among Japanese supermarkets, while their study was limited to the consideration of monopolist behavior. Our study explicitly considers the effect of competition on Japanese supermarkets’ pricing strategies. This study examines the effect of supermarket openings of two national chain stores on incumbents’ pricing of national-brand products based on weekly scanner data compiled by a marketing company that precisely records the name of each product with a scanner bar code, the time of sale, the price, and the number of units sold. The sales information is accompanied with the exact address of the store location. These features of the data enable us to implement the study without paying too much attention to measurement error of the price, timing, and product, which is a major concern in previous studies. In addition, the detailed geographic information enables us to control for unobserved market heterogeneity across regions over time. Stores located within 1.5 km are defined as treatment stores, while those located within a 15-min driving distance but further than 3 km are defined as control stores. This fine definition of treatment and control groups presumably controls for common unobserved local demand shocks. The examination of scanner data clearly shows a drastic price decline between 1999 and 2007 for all six items in the analysis sample: curry paste, bottled tea, instant noodles, instant coffee, detergent, and toothpaste. The analysis results reveal that stores in the treatment group reduce the prices of national brand curry paste, bottled tea, and instant noodles by 0.4–3.1% after the opening of a new supermarket compared with stores in the control group. In contrast, we find very limited or no effects on detergent or toothpaste. The magnitude of incumbents’ price reduction is larger in markets where preexisting conditions were not competitive (i.e., there was only one supermarket in the market area) than markets where it was competitive (i.e., there were three or more supermarkets in the market area). Also, the entry reduces incumbents’ prices only when the entrants are similar in terms of floor area; when general merchandise stores (GMS) enter the market, existing large-scale supermarkets do not reduce their prices. These empirical results are consistent with the theoretical prediction that entry promotes competition in markets where preexisting conditions were not competitive, and the entry is competition-promoting where entrants are close substitutes for incumbents. The rest of the paper is organized as follows. Section 2 describes the data and introduces descriptive statistics. Section 3 explains the empirical method used to identify the causal effect of new supermarket opening on the prices of incumbent stores. Section 4 introduces the basic results and discusses additional results. The last section provides conclusions and proposes possible extensions for future research.

This paper reports evidence on how the entry of new supermarkets in a local market changes the prices of selected national brand items, such as processed food and groceries, at incumbent stores. We have contrasted the price changes of supermarkets that are closely located to the entrant and ones at a distance, based on scanner data with detailed geographic locations of supermarkets. We have found that stores located within 1.5 km reduce prices of curry paste, bottled tea, instant coffee, instant noodles, and toothpaste by 0.9–3.1%. These results suggest that the entry of new supermarkets in a geographic region cuts the market power of incumbent supermarkets and leads them to lower prices. The degree of price reduction is greater when there is only one treatment store, which is consistent with Cournot’s monopolistic competition model. The negative price effects of the new opening are not temporary, but last for a long time, which implies that the welfare gain for consumers is not negligible. We also found that it is large supermarket stores, not GMSs, that bring competition to incumbent food stores. This result suggests that GMSs might have positive spillover effects for an entire region. Although our data contain rich information on prices and entry, to evaluate the total welfare effects of the new entry, we need additional data, such as precise exit information for each store, as well as quantity information for each product at the market level. A data set with a wider coverage of stores would enable us to quantify the consumer surplus. This would be a useful extension of this paper to derive implications for merchandise location policy, as in recent British studies by Griffith and Harmgart (2008). There are many remaining issues to be pursued. In this paper, we have not estimated a structural model of monopolistic competition. To evaluate the welfare effects of the policy change, we need a rigorous structural model, such as those of Salop (1979) and Jia (2008). Other topics to be investigated include the effects on local labor markets, the effects on the productivity of incumbent stores, and regional productivity.