The Phillips Curve Is Not What You Think

This paper examines the empirical foundations and policy implications of the Phillips curve, the canonical inverse relationship between unemployment and inflation in macroeconomic theory. After reviewing the curve's intellectual history and its persistent empirical failures across multiple episodes (the U.S. dot-com boom, the 2007–2009 crisis, the post-Bretton-Woods era overall), the paper argues that the traditional Phillips curve should be replaced by an "exchange-rate Phillips curve." Under this reformulation, the stable negative relationship is not between unemployment and inflation, but between unemployment and real exchange rate appreciation, a relationship that holds across exchange rate regimes. The observed Phillips curve in economies under fixed exchange rates (such as Britain under the gold standard or the U.S. under Bretton Woods) arises as a special case, since inflation and real exchange rate appreciation coincide in such regimes. This reinterpretation has three major policy implications: the relevant trade-off is between unemployment and external competitiveness rather than inflation; the inflation rate is not a reliable guide for macroeconomic stabilisation; and the concept of a natural rate of unemployment (NAIRU) loses its empirical grounding.