One of the aspects of Milton Friedman’s work which won him the Nobel Prize was “his demonstration of the complexity of stabilization policy.” We consider the question of what that meant, suggesting that his early 1950s “formal analysis” of the effects of full-employment policy was his most important contribution in the area. We address the question of the relation of that piece to his other thinking, particularly on the length and variability of lags in the economic system. Since 1976 the “formal analysis” has been relatively little noted, and we suggest that the prominence given to stories about the Phillips curve are part of the reason, and that another part is that in making the case for policy rules Friedman himself downplayed the formal aspect of his thinking in favour of a more ad hoc empirical one.