Changing Market Structure and Evolving Ways to Compete: Evidence from Retail Gasoline.

AuthorKim, Taehwan
  1. INTRODUCTION

    Firms constantly innovate, and innovations force changes in the competitive landscape. Firm innovations have been remarkable, particularly in the retail sector, and the introduction of automation is one example of firm innovations, altering the optimal allocation of resources of buyers and sellers (Bronnenberg and Ellickson, 2015 and Hortagsu and Syverson, 2015). There is a growing body of research studying the employment effects of automation, but relatively little quantitative evidence on how automation impacts prices of traditional and incumbent firms. (1) In this paper, I use a case-study approach to examine the competitive effects of an innovation in a retail-gasoline market and jointly investigate how an innovation changes ways firms compete on price: the introduction of self-service technology in retail gasoline in Korea.

    The retail-gasoline industry is useful for examining the pricing behavior of sellers because the product is almost homogeneous and prices are clearly posted and observed by consumers. My setting is unique in that I have high-frequency, station-level data starting in 2010, when self-service sellers were very rare, and continuing through 2015, when they accounted for a quarter of the market. I study sellers' pricing by service level, and find that gas stations that continue to offer full-service differentiated themselves by offering bundled products and services and raised their prices during this period.

    Specifically, the price gap between full-service and self-service increased most dramatically in Seoul; the full-service premium was 2% in early 2010 and increased to 8% by 2015. (2) At the same time, the distribution of full-service prices became increasingly right-skewed while the variability of self-service prices remained relatively stable. These stylized facts clearly suggest a different evolution of sellers' pricing strategies by service level during the market transition from full-service to self-service.

    My empirical findings are three-fold. First, I use the prediction of monopolistic competition models--greater competition lowers price--to test whether the exit of full-service stations can explain the increase in the full-service premium. (3) If the gasoline market is segmented by service level, then the increase in the full-service premium could be explained by the falling number of full-service stations and the increasing number of self-service stations. However, in a difference-in-difference (DID) specification, the changes in the service composition of gas stations barely explain the increasing full-service premium in this market, and the trend of full-service premium is very robust to measures of the competitive environment.

    My second finding is a confirmation that self-service stations not only offer competitive prices but also drive down prices of nearby competitors. DID specifications show that, ceteris paribus, self-service stations charged 5% less per gallon on average during my study period. Regarding competitive effects, price elasticity with respect to competition is much higher in absolute value when competitors are self-service than when they are full-service. Specifically, having one more self-service station within a 1-mile radius is associated with a 0.5% reduction in gas prices, whereas entry of full-service stations has no statistically significant effect on prices. This finding implies that self-service stations compete for price-sensitive customers who may prefer only a low-priced gasoline component to a bundle of gasoline and service.

    The third finding includes evidence of full-service stations' subtle differentiation on one or more non-price dimensions. For example, some gas stations combine with other types of businesses, including dry cleaners, nail salons, or fast-food restaurants, and some stations offer extra products free of charge, such as coffee, a carwash, or vacation packages, depending on the amount of gasoline purchased. (4) These unique features are generally provided by full-service stations whose gasoline prices are significantly higher than the average price for full-service. It is my claim that such product repositioning is the new strategic choice of high-cost marginal sellers in response to the emergence of low-price competitors.

    To supplement the descriptive evidence, I perform a novel test using my price panel data. That is to compare stations that bundle products to those that do not. Using information on stations' bundling choices in May 2017, I divide stations in my price panel (from 2010 to 2015) into two groups: stations that later bundle products vs. those that do not. In doing so, I trace the price difference between later-bundling stations and later-non-bundling stations. I find that the price difference between the two groups increases in the later period of my sample, consistent with stations increasingly bundling their products and charging a premium during the market transition.

    Differentiated sellers enjoy less intense competition on price, which allows them to charge high and stable prices. Consistent with this implication, my descriptive evidence and quantitative analysis together support the claim that sellers' strategic choices evolve in different ways; higher-priced stations compete for less-price-sensitive consumers, while lower-priced stations that usually offer self-serve gasoline focus on price-sensitive consumers. Taken together, these features have led to an increase in the full-service premium during the market conversion from full-service to self-service.

    This paper contributes to a broad literature on gasoline pricing and its softening effect on competition, such as Shepard (1991), Png and Reitman (1994), Contin, Correlje, and Huerta (1999), Lewis (2008) and more recently Soetevent and Bruzikas (2018). (5) Shepard (1991) is the first study to document price discrimination by gas stations and estimate the premium charged for full-service gasoline. Png and Reitman (1994) focus on service competition as one aspect of station quality and estimate the premium for a service-time reduction at gas stations. Contin, Correlje, and Huerta (1999) examine competitive effects on price after price deregulation in the Spanish gasoline market and find a price reduction driven by new entrants. Lewis (2008) explores the gasoline market in San Diego on measuring price dispersion with differentiated sellers on a brand dimension.

    Most related to this study is Soetevent and Bruzikas (2018), who exploit the transition of the Netherlands retail-gasoline market from self-service staffed stations to fully automated stations. They estimate, by employing a DID matching strategy, that prices around automated stations decrease on average 0.2% through competition. They also argue that prices at the technology adopted stations immediately reduce but the insignificant difference in impact by the timing of the process innovation. With regard to the market transition, the setting of this paper is similar to that of Soetevent and Bruzikas (2018), but the main interest here is to examine the pricing behavior of high-cost sellers that do not adopt new technology. Furthermore, my paper is the first in retail-gasoline literature to document that product differentiation can be observed in a more subtle form, different from simply location and brand.

    Several other papers have examined the relationship between product differentiation and its softening effect in a retail market other than gasoline. This study is also in line with these studies. For example, Grabowski and Vernon (1992) examine competitive effects of generic drugs on price and find that the entry of generic drugs causes the price of major branded drugs to increase. Mazzeo (2002) examines motel markets located along U.S. interstate highways, and finds that the effect of competition on price is little when motels are differentiated. Basker and Noel (2009) show that the effect of Wal-Mart's entry on competitors' prices is greatest at low-end chains that compete for price-elastic consumers, and smaller at supermarkets that differentiate themselves from Wal-Mart.

    The rest of the paper is organized as follows. The evolution of market segmentation is described in Section 2, and Section 3 explains the data used in this paper. Section 4 presents my main results on the pricing behavior of sellers by product position. Section 5 concludes.

  2. The Evolution of Market Segmentation

    2.1 Full-Service vs. Self-Service

    The first introduction of self-service technologies in the Korean retail-gasoline market was in 1993, but self-service stations failed to attract customers and soon disappeared. A gas station with self-service pumps opened again in 2003, but the self-service format was still rare until the end of 2007, when it accounted for approximately 0.3% of the market. The market conversion from full-service to self-service accelerated in early 2008, possibly due to a sharp increase in global oil prices. New self-service stations have opened every month since then. Self-service stations constituted 17% of all stations as of December 2015 at the country level, and most were converted from full-service stations rather than being new entries.

    The distinction between full-service and self-service is whether consumers must place their labor into the production function of stations. (6) Consumers who purchase self-service gasoline fill their gas tank and pay at the pump themselves, whereas those buying full-service gasoline wait for a service attendant in their car and request the amount of gasoline they want to buy. In the early 2000s, a few full-service stations in Korea used to provide "full services" such as cleaning the windshield and checking under the hood of a car, similar to the U.S. market in the 1960s. However, such services have disappeared. (7) In Korea, full-service consumers nowadays expect to receive only pumping and paying services, or at most, a...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT