Revisiting the Phillips Curve: The Empirical Relationship Yet to be Validated*

Abstract

The Phillips curve began life in 1958 as a simple curve-fitted relationship between the rates of wage inflation and unemployment and went on to play a crucial role in the broader development of macroeconomics, giving rise to several controversies about its interpretation and role in policy-making. Recently, the traditional narrative about its theoretical underpinnings has been called into question as a sequence of ‘stories’ to provide support for particular theoretical perspectives on macroeconomics. The primary aim of this paper is to challenge the conventional wisdom relating to the Phillips curve being an attested empirical relationship, by showing that the empirical evidence of the most influential papers that helped to frame the traditional narrative is untrustworthy, in the sense that the probabilistic assumptions invoked by their inferences are invalid. That is, not only the traditional theory-driven narrative is misleading, but the empirical evidence used to corroborate it is untrustworthy.