This paper estimates a structural model of the Brazilian carbonated soft drink industry to test the claim that the observed low prices of low-end entrants owe to marginal cost advantages over the large, established brands, allegedly stemming chiefly from tax evasion. Such entrants, numbering in the hundreds, are typically small-scale operations, with limited geographic reach and no advertising. In addition to the low-cost hypothesis, advocated by the incumbent duopolists, the model allows for other (complementary or substitute) explanations: that consumers have different preferences for the low-price entrants over the established brands, and that firm-level strategic behavior is heterogeneous. The paper draws on a rich original panel dataset to structurally inform the relative weight of each hypothesis in explaining the observed price differences. The paper finds some support for the low-cost hypothesis, but finds strong support for the demand side hypothesis: the established brands' market power almost single-handedly explains the price premium they command over the entrants. It provides an innovative application of structural IO modeling and estimation within the realms of international business strategy, public finance, and development.
Structural IO estimation demand estimation business practices in developing countries firm-level heterogeneity informal economy