In the now conventional view of the inflation process, the New Keynesian Phillips Curve (NKPC) captures most of the persistence in inflation. The sources of persistence are twofold. First, the ?driving process? for inflation?the output gap or, more commonly, real marginal cost?is itself quite persistent, and a casual inspection of the NKPC reveals that inflation must ?inherit? this persistence. Second, a modest amount of backward-looking or indexing behavior imparts some ?intrinsic? persistence to inflation. This latter source is generally thought to be of less importance than the former, as the degree of autocorrelation in the driving processes is substantial. This paper shows that in practice inflation in the NKPC inherits very little of the persistence of the driving process, and, contrary to conventional wisdom, it is intrinsic persistence that constitutes the dominant source of persistence. The paper explores the reasons for this and links them to two empirical observations. First, it has been difficult to develop a sizable coefficient on the driving process in NKPCs. Second, the shock that enters the NKPC, while often difficult to motivate economically, is large and is critical in distinguishing the sources of inflation persistence. While these observations help to clarify the behavior of inflation in NKPCs, they raise other fundamental questions about how to model inflation.

This paper was revised in September 2005.