Allies or rivals? Spatial price competition in the Chinese retail gasoline market of inner Mongolia

ABSTRACT

This study examines the pricing behaviour in the gasoline market using station-level pricing data from Inner Mongolia, China. The spatial econometric findings confirm that: (1) there is a significant spatial dependence among gas stations; (2) the two state-owned oil giants, PetroChina and Sinopec, receive a positive pricing effect from their brands compared with other stations; (3) there is no significant evidence of price competition among PetroChina stations, and some evidence of competition among Sinopec stations, and more intense competition among other stations; and (4) ‘stand-alone’ (no stations nearby) stations have a price premium, but the premium is limited by the price regulation. Our results show the heterogeneity in pricing behaviour between stations, implying the need for differentiated policies to ensure orderly competition in the retail gasoline market.

KEYWORDS:

• price competition
• spatial models
• retail gasoline
• China

JEL:

• C23
• L13
• L81
• R32

ACKNOWLEDGEMENTS

The authors appreciate the valuable comments of Professor Paul Elhorst (editor-in-chief) and two anonymous referees that helped to improve the paper significantly.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Notes

1 See https://www.caixinglobal.com/2018-06-30/china-removes-limits-on-foreign-ownership-of-gas-stations-101288799.html/.

2 Three types of gasoline are offered in most gas stations in China: RON 92 (#92, regular), #95 (premium) and #98 (premium). The numbers 92, 95 and 98 represent the research octane number of the gasoline. Octane number (ON) is a measure of the ignition quality or flammability of gasoline, which include research octane number (RON) and motor octane number (MON). In China, the octane rating on the pump is simply the RON figure. The United States, by contrast, uses the arithmetic average of RON and MON, called the anti-knock index (AKI).

3 The setting of Alone does not mean that we simplify the predetermined W matrix, which is used to analyse whether there is heterogeneity in the marginal effects of the explanatory variables for gas stations within and outside of 3 km.

4 When studying other forms of competition, the negative alone can also be used to determine the existence of competition (Griffith & Arbia, 2010; Wang, 2018). For instance, Griffith and Arbia (2010) examined the geographical competition for land surface. When the area is limited, areal expansion of a given unit results in size reductions of its neighbouring units, creating geographical competition for land surface. Therefore, a negative indicates competition. In a recent study, Wang (2018) focused on the price deviations rather than price levels of internet book listing. A key characteristic of online book sales is that sellers usually have little information about the demand side, hence their pricing decisions do not necessarily have a strong supply–demand-type structural foundation. Therefore, the negative in his study implies that there is price dispersion, which is the result of predatory pricing. However, in the gasoline retail market, because the information is relatively perfect, it conforms to the classic supply–demand analysis, and is often studied with the Bertrand model (Lee, 2009; Pennerstorfer, 2009; Alderighi & Baudino, 2015; LeSage et al., 2017).

5 Only one observation in the sample has a price that exceeds the price ceiling, which is due to insufficient supervision.

6 Indeed, some anecdotal evidence also indicates that private gas stations provide poor-quality gas (http://www.xinhuanet.com/2019-08/13/c_1124868786.htm).

7 This study only explains the mechanism of ally share’s negative impact on price from the perspective of market competition. Other mechanisms could explain this response. For example, the space gathering of the same brand will save transportation costs and reduce prices. However, due to the lack of cost data, unfortunately this study cannot test the mechanism behind it. Many thanks to the reviewer for raising such an issue.

8 Indeed, to examine the impact of a leading or dominant gas station in a rigorous manner, it is theoretically possible to specify the spatial econometric model as that in Holly et al. (2011) and Yesilyurt and Elhorst (2017), and empirically test (say, via the LR test) whether a particular gas station is dominant unit, although dominant units have not been prevalent in the spatial econometrics literature (Elhorst et al., 2021) mainly due to that the setting, estimation and testing of such models are more complicated than traditional spatial econometrics.

Additional information

Funding

The authors appreciate the financial support of the Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China [grant number 21XNH191].