10 October 2017
According to preliminary estimates, non-farm payroll employment fell 33k in September, an outturn that would ordinarily put policymakers and markets into a spin. However, this was no ordinary month. Thanks to hurricanes Harvey and Irma, almost 1.5 million people were estimated to have been unable to get to work during the survey week, significantly distorting the figures. Indeed, when we take this and the other components of the report less distorted by weather effects, the message was much more bullish, and reinforces the picture of an economy growing faster than its potential. Both headline unemployment (4.2%) and under-unemployment (8.3%) rates fell to new cycle lows, with the former declining to its lowest level since December 2000. This progress came in spite of an increase in the participation rate, reinforcing the nascent signs that discouraged workers are gradually being drawn back into the labour force. Taken together, declining unemployment and rising participation lifted the employment-to-population ratio by 0.3 percentage points (ppts), its highest level since January 2009 when the fallout from the financial crisis was still gathering force.
The continued rise of labour utilisation rates begs the question of whether labour market slack has now been eliminated. For example, Federal Open Market Committee members currently think the long-run equilibrium unemployment rate is somewhere between 4.4% and 5%, suggesting that the economy is already at, or beyond, full employment. OECD and Congressional Budget Office staff also think the structural unemployment rate lies within this range. Other indicators, however, imply that some slack may still exist. The U6 measure of labour underutilisation, which takes into account marginally attached workers and those working part-time for economic reasons, fell as low as 7.9% in the last cycle, while at 78.9%, the employment ratio of prime-aged workers (those aged 25-54) is still 1.5 ppts below the peak of the previous cycle (see Chart 2). Some of that shortfall probably reflects structural factors, including the impact of the opioid crisis, but how much is still highly uncertain.
The typical way to resolve questions around slack is to examine the trajectory of wages and unit labour costs. Those in the camp that slack has been eliminated would point to the fact that average hourly earnings increased 0.5% in the month of September, sending the year-on-year growth rate to 2.9%, its strongest since June 2009. Yet these data too may have been distorted by hurricanes and wage growth of private sector production and non-supervisory workers still lies within its range of the past two years. Meanwhile, there is little evidence of accelerating wage growth in the employment cost index, or the Atlanta Fed’s wage growth tracker, and unit labour cost growth has actually been decelerating (see Chart 3). The transmission of changes in labour market slack into labour cost measures is of course long and variable, but there is not an open and shut case that the economy is running hot and that a wage and inflation breakout is imminent. The upshot is that although the Federal Reserve has more than enough ammunition to justify lifting the federal funds rate again in December, we still think that underlying inflationary pressures are likely to build only gradually and hence the pace of monetary policy withdrawal will remain very gradual.