A model of the shareholder constraint is described in terms of the relationship between shareholding concentration and corporate control. A unified perspective (including takeovers as a special case) is developed whereby leading coalitions are costly to form and possess power in a shareholder-voting game. Control is defined in terms of power indices for simple games. A static theory of the firm is developed in terms of the optimal formation of controlling coalitions. A dynamic theory of an owner-controlled firm is described which explains increasing shareholding dispersion accompanying growth.