The Phillips Curve Is Alive and Well

Article excerpt

The Federal Reserve Bank of Boston's 1978 economic conference was entitled "After the Phillips Curve." Many of the papers in the conference volume sounded the death knell of the Phillips curve, citing its dismal performance in the face of oil shocks in the 1970s, and its inappropriateness as a policy guide because of its presumed sensitivity to shifts in the underlying macroeconomic structure.

Seventeen years after the publication of the conference volume, rumors of the death of the Phillips curve appear to have been greatly exaggerated. In fact, the Phillips curve is alive and well, and living in a good number of (although certainly not all) widely used macroeconometric models. This paper takes the view that the primary reason for its longevity is that, in contrast to the common perception at the time of the 1978 conference, the Phillips curve has been an extremely robust empirical relationship, showing little or no sign of instability over the past 35 years. To outward appearances, at least, the Phillips curve is as structural a relationship as macroeconomists have ever had at their disposal.

The major criticism levied against the Phillips curve and many macroeconometric models of the 1960s and 1970s was that they were not truly structural. That is, they captured empirical regularities between aggregate variables like the unemployment rate and the rate of inflation, but they did not take account of all the interactions in the underlying or structural behavior of consumers and firms in the economy. The risk to this approach, as articulated by Lucas and Sargent (1978), was that even if the underlying structural behavior of economic agents remained stable, the measured relationships among aggregate variables could easily shift as conditions changed in the linkages not incorporated into the aggregate relationships. A fuller description of this criticism (the "Lucas critique") and a simple example are described in Section III below.(1)

Ultimately, however, as Lucas and Sargent (1978) emphasized, "the question of whether a particular model is structural is an empirical, not a theoretical, one." Thus, while the theoretical force of Lucas and Sargent's arguments still holds--macroeconomic models not based on underlying structure are subject to shifts--it is the empirical force of their point that is of interest for macroeconomists and policymakers. This paper examines an array of empirical evidence bearing on the stability of the Phillips curve, and calls into question the empirical force of the Lucas critique as it applies to the Phillips curve.

I. A Brief History of the Phillips Curve

The essence of the modern Phillips curve is that the rate of change of nominal wages depends upon the expected rate of change of the overall price level--workers want their wages to keep pace with inflation, all else equal--and on the level of the unemployment rate relative to its natural rate. The dependence of wage inflation on expected inflation (as well as unemployment) is the difference between the "expectations-augmented" Phillips curve, first implemented empirically in the 1960s, and the original Phillips curve.(2) The natural rate of unemployment, or NAIRU (non-accelerating-inflation rate of unemployment), is defined by the behavior of inflation: It is the rate of unemployment that exerts neither downward nor upward pressure on wage inflation, given expectations of price inflation.(3)

Many current versions of the Phillips curve cast it as the link between the rate of inflation in overall prices and the level of unemployment. In fact, this portrayal is a simplification of an underlying relationship between the rate of change of wages and unemployment, and among the levels of wages, prices, and productivity. The first link in this chain is the original Phillips curve; the second link relates prices to unit labor costs, defined as the difference between wages and productivity.(4)

The strong empirical bond between the rate of change of wages and the level of unemployment was first documented in Phillips (1958) for United Kingdom data from 1861 to 1957. …