DiMaggio and Powell (1983) argued that organizations, in their quest for legitimacy, are subjected to isomorphic pressures which produce increasing similarity among peer organizations over time: "Once an organizational field becomes well established ... there is an inexorable push toward homogenization." Yet, in contradiction to this "iron cage" hypothesis, many industries became more heterogeneous, not more homogeneous, in their profiles during the latter decades of the twentieth century, particularly between about 1980 and 2000 (at least on the American landscape). Why didn't "inexorable homogenization" occur? We argue that DiMaggio and Powell were correct about the forces that give rise to isomorphism but failed to anticipate several major macrosocial trends that caused those forces all to move in directions that diminished, rather than accentuated, isomorphism. For example, DiMaggio and Powell argued that ambiguity about goals will propel isomorphic change; but the goals for publicly-traded U.S. corporations became less ambiguous. They hypothesized that the fewer the alternative organizational models in a field, the faster the rate of isomorphism; but the array of organizational models increased significantly. We empirically illustrate the increased heterogeneity that occurred within American industries by tracing the trend toward divergence - on several dimensions of strategy and performance - within the steel industry. An analysis of 18 additional industries similarly yields far more evidence of increased heterogeneity than of increased homogeneity over the latter decades of the twentieth century. We go on to argue that reduced isomorphic pressures not only engendered greater intraindustry variety, but also increased managerial discretion, which contributed greatly to the romanticization of CEOs that occurred during the period 1980-2000.