Under Ms. Yellen’s leadership, the preemption of inflation based on the Phillips curve has become the lodestar of monetary policy. Accordingly, last week’s Fed statement, while allowing that inflation “is expected to remain somewhat below 2 percent in the near term,” reaffirms the view that it will “stabilize around the Committee's 2 percent objective over the medium term.”
The Phillips curve is a graphical representation of the ostensible inverse relationship between inflation and unemployment, implying that it should be downward-sloping. Yet, following the post-recession rebound in inflation in 2009, as Chart 1 shows, the relationship between the unemployment rate (horizontal axis) and the PCE inflation measure favored by the Fed (vertical axis) has been nothing of the sort.
Each dot on the chart represents the combination of the jobless rate and the inflation rate seen in every month since the end of 2009. While there is no obvious pattern in the data, the best-fit line is actually upward-sloping, as evident from the chart.
In other words, the Phillips curve is badly broken. Yet, this should not be a novel insight.
As we wrote at the end of the last century, “[t]he ‘Phillips curve’ relationship between unemployment and inflation has long been a dubious proposition. The Future Inflation Gauge remains a better guide to the direction of inflation” (U.S. Cyclical Outlook, December 1999). That 20th-century insight has remained valid in the years since the financial crisis that have so baffled mainstream economists, including Ms. Yellen.
Chart 2 shows the exact same data points as Chart 1, the difference being that they are now color-coded; and inflation cycle (IC) peak and trough dates, depicted by larger diamonds, have also been added to the chart. Monthly data falling within IC downturns are marked by blue diamonds, and data covered by IC upturns are marked by red diamonds. The corresponding red and blue arrows indicate IC upturns and downturns, respectively.
The patterns in the data now become self-evident. Regardless of the unemployment rate, inflation falls during IC downturns (blue arrows) and rises during IC upturns (red arrows). Of course, ECRI’s U.S. Future Inflation Gauge (USFIG) is designed specifically to anticipate these IC downturns and upturns, and is therefore “a better guide to the direction of inflation” than the Phillips curve or its variants.
The USFIG is in a cyclical downturn. Following that downswing, a fresh IC downturn has begun (leftmost blue arrow), which is why inflation, as measured by PCE deflator growth, is falling. Given the downturn in the USFIG, it will continue to decline in the coming months, regardless of the decline in the unemployment rate below “full employment” this year.
Eventually, of course, the inflation cycle will turn back up. But it is the USFIG – not the Phillips curve – that will provide early warning of that directional shift.