US Productivity - Out of the Slow Lane?

Most economists and policymakers believe that since around 2005 the US has been stuck in a 'low productivity growth' regime in which labour productivity oscillates around an average of 1.33% yoy, compared with the previous 'high productivity growth' regime which held from early 1997 to late 2004 in which productivity averaged 2.9% a year.

The assumption that there’s a ceiling on US labour productivity growth, coupled with an assumption of a demographically-constricted ceiling on employment growth feeds into calculations about sustainable growth rates. And those calculations will go a long way to determining how the FOMC feels obliged to act.

But the low-productivity assumption is under challenge now, with recent readings knocking up against the ceiling: in 2Q labour productivity rose an annualized 3%, and the first-estimate of 3Q this week came in at an annualized 2.2%.

Taking real GDP per worker as a crude but easily calculable measure of productivity, the 1997 to 2004 average comes out at 2.2% with a standard deviation of 0.8%, whilst the 2005-current average comes in at 0.9% with a standard deviation of 1%. If we exclude the exceptional 'recovery period' of 4Q09 to 3Q10 (when this measure of productivity growth averaged 3.5%), then the low regime of productivity growth averages 0.7%, with a standard deviation of 0.7%.

If we now consider the last four quarters performance (4Q17 1%, 1Q18 1%, 2Q18 1.2%, 3Q18 1.3%), we can see that this is now consistently pushing towards the ceiling of this 'low productivity' regime.

productivity regimes.jpg

There are two possible interpretations of this, and both are in different ways destabilizing. One view is that there is no reason to expect the low-productivity regime to be challenged. In which case, we must expect to see the performance of the last two quarters as exceptional peak-cycle phenomena which will be swiftly followed by a cyclical decline. This is as good as it gets.

The alternative view is that there is nothing sacrosanct about the 'low productivity regime' and that the gains of the last two quarters are presaging a reversal back to a higher productivity regime.

What view you, and policymakers, take of this is absolutely critical, because it will be one of the assumptions which determine the possible sustainable growth trajectory of the US, and consequently what is the 'natural rate of interest' around which the Fed will attempt to guide the economy.

Despite the ambiguous/ambivalent structure of 3Q’s 3.5% GDP growth, I think there is a good reason to believe that the US is transitioning to a higher productivity regime. That good reason is to do with how capital stock per worker helps determine labour productivity growth.

As the chart below shows, between 2010 and late 2016, growth in labour productivity was constrained by negative or historically very weak growth in capital stock per worker. In fact, when one stripped out the impact of capital per worker, the performance of output per worker was markedly better than in the so-called 'high labour productivity' period.

Since 2016, growth in capital per worker has accelerated back towards the lower boundary of pre-crisis normality, but with surprisingly little deterioration in output per worker less capital per worker. Even into 3Q18, this measure remains noticeably better than what was usually achieved prior to the crisis. By itself, this chart gives no reason at all to expect any deterioration in overall output per worker in the near future. In other words, no reason to discount the exit from the post-2004 'low productivity regime'.

elements of productivity.jpg