Abstract

We extend the standard supply side model of an imperfectly competitive, differentiated-product industry to include licensing between competitors and derive the implied Bertrand-Nash equilibrium conditions. The model permits a novel empirical analysis on the role of licensing for genetically engineered (GE) traits in the US corn and soybean seed industry. A discrete-choice framework motivates the specification of a two-level nested-logit seed demand model, which we estimate using a large dataset of US farmers’ corn and soybean seed purchases over the period 1996-2016. The estimated model enables us to recover firms’ marginal costs and the implied royalty rates for GE traits. The analysis provides new insights on the exercise of market power in the context of an input market that is critical to innovation in agriculture. Estimated markups over marginal cost are sizeable, and royalties for GE traits contribute a non-trivial amount to these markups. Still, the extent of market power exercised in the baseline Bertrand-Nash equilibrium is much lower than what would result from full collusion. Notwithstanding its strategic effect on pricing, licensing of GE traits to competitors actually has net positive impacts for all market participants, including seed buyers (farmers).