Exclusion or Subsidization? A Competitive Analysis of Quality Regulation Strategy for Two-Sided Platforms

Abstract

With two-sided platforms becoming an increasingly ubiquitous business model, quality is a vital factor for the success of high-technology platforms that face fierce competition. To maintain competency, high-technology platforms commonly use two quality regulation strategies: the exclusion strategy (E strategy), in which the platform denies access to low-quality complementors, and the subsidization strategy (S strategy), in which the platform provides a fixed subsidy to high-quality complementors. This paper investigates the optimal quality regulation strategy for platforms in a duopoly setting. We examine and compare three scenarios: (i) both platforms adopt the exclusion strategy, i.e. mode EE, (ii) both platforms adopt the subsidization strategy, i.e. mode SS, and (ii) one platform adopts the subsidization strategy while the other adopts the exclusion strategy, i.e. mode SE. First, we find that although the developer network size is larger and the platforms charge developers higher access fees under SS, the average quality and the consumer access fees are lower under SS than under EE, leading to lower profits for platforms. Second, under SE, in comparison with the platform that adopts the exclusion strategy, the platform that uses the subsidization strategy achieves lower average quality and larger network sizes on both sides but may set higher or lower access fees on both sides. Moreover, the platform under the subsidization strategy profits more (less) when the operation cost on the developer side is high (low). Third, asymmetric mode SE does not necessarily induce moderate outcomes for market participants compared to modes EE and SS. We also examine the equilibrium mode by considering platforms’ optimal strategies for quality regulation. Our analyses reveal that as the operation cost on the developer side increases, the equilibrium mode evolves from EE to SE/ES and then to SS. These results and insights are robust to several alternative assumptions.