This paper makes two straightforward points that we argue are central to understanding the literature and debate surrounding the stability of the Phillips curve. First, the endogeneity of monetary policy implies that aggregate data are largely uninformative as to the existence of a stable relationship between unemployment and future inflation. Second, if the NAIRU model is assumed to be true, regional data can be used to identify the structural relationship between unemployment and future inflation. We find that a 1 percentage point increase in the unemployment rate is associated with a roughly 0.3 percentage point decline in inflation over the next year.