Is this expansion’s weak productivity growth a “statistical illusion?”
Recent better productivity likely has legs heading into 2018.
Stocks have done best following periods of weak productivity.
My last report was on the acceleration in business capital spending (capex) that is likely to be an economic highlight in 2018. Part-and-parcel of capex is productivity—officially known as non-farm labor productivity—which has averaged less than 1% annualized growth during the current expansion. This has been the weakest era of productivity for any expansion in modern history, with the exception of the late-1970s. But all is not lost. Not only do I think productivity is at least a bit understated—based on how it’s calculated—but there’s reason for optimism for an uptick in 2018.
As noted by GavekalDragonomics in a great report a couple of years ago, the apparent productivity slowdown is a “statistical illusion” caused by rapid technological change. The analysis starts with a famous remark made in 1987 by economics Nobel Laureate Robert Solow: “We can see computers everywhere except in the productivity statistics.” In other words, traditional data calculations are less reliable in the midst of rapid structural and technological changes, especially when a large and expanding proportion of employment/consumption growth is generated by the services side of the economy. This is decidedly different than the manufacturing-oriented era when non-farm productivity calculations were first developed.
As a result, many ‘old economy’ activities have been replaced by online services operating on totally different business models and often delivered free at the point of sale. Such businesses are underestimated or even missed completely in gross domestic product (GDP) and productivity statistics, even though they provide services of huge economic value. The measurement problem arises because many online services generate far less revenue than the old economy businesses they replace, yet they employ large workforces and pay high wages.
The fact that countries with the fastest business innovation are also the ones apparently suffering the largest productivity disappointments corroborates the conjecture that the fault lies not in businesses’ ability to innovate, but in statisticians’ ability to measure the true value of that innovation.
There is appropriate enthusiasm about the impact of technological advances like robotics and artificial intelligence (AI); which could lead to more rapid productivity growth. Quality enhancements and innovations tend to occur more rapidly in the digital economy. There is also reason for hope using traditional productivity measurements; along with productivity’s leading indicators.
We are already witnessing a sharp rebound in productivity this year. After a drop into negative territory in 2016, the year-over-year change moved back up to 1.5%; while it’s running at a 3% annualized rate as of the third quarter. On a five-year percentage change basis, we are likely stabilizing, as seen in the chart below. Assuming further acceleration, this is good news for wages as the correlation between productivity and wage growth has historically been tight.