Cannibalization, Innovation and Spin-outs

When an idea for a new product arrives, will it be developed and by whom? We develop a spatial model in which an idea arrives to a researcher within the firm. Products are imperfectly substitutable, so that developing a new product that is close to an existing product will cannibalize some amount of the existing product's sales, and the cost to develop a new product is higher the further it is from an existing product. Together these forces mean that there exist ideas that can be developed more efficiently by the researcher as a spin-out than by the firm (due to the cost of fit) but that the firm prefers to buy out the researcher and either develop itself or discard (due to the potential loss from cannibalization). These inefficient outcomes occur for ideas at intermediate distance from the firm's existing portfolio, and are likelier and more severe the higher is demand and the greater the degree of substitutability.