Many positive and normative results in models using the monopolistic competition framework of the Dixit-Stiglitz-Spence type depend on the value of an important parameter, the 'taste for variety'. In these models, however, the taste for variety is linked in a close but arbitrary manner to the degree of market power of monopolistic competitors. In this paper we make a simple modification to the basic model, which clearly disentangles taste for variety and market power. We then show how this changes the usual comparisons between social optimum and market outcomes.