Ronald Coase's celebrated query as to why economic actors typically aggregate into entities called ‘firms’ rather than transacting as individuals in a market has engendered a vigorous stream of research. This paper asks a parallel question: why is it that in all modern economies, firms themselves aggregate into larger entities, often more stable than any literature predicts, which are here referred to as ‘business groups’? After establishing some working definitions, and discussing the curious conjunction of empirical importance and analytical invisibility of business groups, an attempt is made to establish the most significant dimensions along which such groups vary. We end with some speculations on the role of these groups in economic development.