The new orthodoxy in studies of policy dynamics is that policy change occurs through a homeostatic process. “Perturbations” occurring outside of an institutionalized policy subsystem, often characterized as some type of societal or political upheaval or learning, are critical for explaining the development of profound and durable policy changes which are otherwise limited by “endogenous” institutional stability. These homeostatic assumptions, while useful for assessing many cases of policy change, do not adequately capture the historical patterns of policy development found in many sectors. The roots of this problem are traced back to the origins of the new orthodoxy in comparative policy research whereby different levels (orders) of policy making have been incorrectly juxtaposed, providing a parsimonious, but sometimes empirically incorrect, view of policy change. Revising existing taxonomies of policy levels provides a superior identification of the processes of change, and uncovers more than one mechanism through which significant policy change can occur. Three of these alternative mechanisms – a “neo-homeostatic” one in which paradigmatic changes occur through endogenous shifts in goals; a “quasi-homeostatic” in which exogenous factors influence changes in objectives and settings; and a “thermostatic” one in which durable policy objectives require that settings adapt to exogenous changes – are discussed.