A BLP Demand Model of Product-Level Market Shares with Complementarity

Applied researchers most often estimate the demand for differentiated products assuming that at one item can be purchased. Yet simultaneous multiple purchases are pervasive. Ignoring interdependence among lead to erroneous counterfactuals, in particular, because complementarities ruled out. I consider identification and estimation of a random coefficient discrete choice model bundles, namely sets products, when only product-level market shares available. This last feature arises aggregate as opposed individual available, very common phenomenon practice. Following classical approach with data, two-step method. First, using novel inversion result which exhibit Hicksian complementarity, recover mean utilities from shares. Second, infer structural parameters while dealing price endogeneity, use instrumental variables. propose practically useful GMM estimator whose implementation is straightforward, essentially standard BLP estimator. Finally, Ready-To-Eat (RTE) cereals milk US. The estimates suggest RTE overall complementary synergy consumption crucially depends on their characteristics. such results misleading counterfactuals.