Retail pricing outcomes in a dual-channel supply chain: Studies from the petroleum industry

This dissertation contains three essays relating to vertical integration in the petroleum industry and its relationship to retail gasoline prices. The first essay of the dissertation is a policy paper about price differences between partially integrated lessee operated stations and wholly integrated company operated stations and acts as a good introduction for the other two essays. The second essay employs structural estimation to determine the relative importance of various differences in price setting incentives. The third essay looks at how different stations impact pricing dynamics. There has been some debate about the real world impact of vertical integration on prices for consumers. In the retail gasoline market this has taken the form of debate over the efficacy of divorcement legislation prohibiting major oil companies from integrating into the retail market in the form of company operated stations. In the first chapter, we supplement existing literature studying divorcement by examining daily station-specific pricing differences. We find that vertically-integrated company operated stations charge prices roughly 5 cents lower than partially-integrated lessee operated stations. The second chapter explores the importance of three ways in which the pricing decisions of integrated company-operated retailers differ from the pricing decisions of solitary non-integrated retailers. The first reflects double marginalization wherein the non-integrated retailer charges a higher price because the combined markups of the wholesaler and the retailer exceed the markup of the integrated retailer. The second difference is a multimarket brand pricing reputation mediated by consumer search or switching costs which is important to the integrated firm operating in multiple markets but not to the solitary non-integrated retailer. The third difference reflects the concern of price setters at integrated retailers for stealing demand from other stations of the brand if consumers substitute between them. We construct a dataset for a regional gasoline market in order to explore the magnitude of each effect. The Arellano-Bond method for a dynamic panel was used to estimate demand parameters and two-stage least squares was used to estimate conjectural derivatives. Results indicate that the multimarket brand pricing reputation was a significant determinant of higher pricing by the non-integrated retailer. Existing research has identified several interesting pricing dynamics in the gasoline market. The third chapter extends this literature by examining the effects of having two types of operators in the market - stations with prices set by an integrated company-operator and stations with prices set by a non-integrated lessee-operator. With heterogeneous station operators, we are able to not only reproduce existing empirical findings, but also identify several new empirical regularities. For instance, by linking data on wholesale prices charged to the lessee-operated stations, we identify how heterogeneity in costs between company-operators and lessee-operators can lead to asymmetry in price adjustments and to price leadership. Simulation results based on a model that incorporates heterogeneous station operations confirms the ability of such a model to replicate key pricing dynamics that we find in our data