The productivity slowdown does not result primarily from the mismeasurement of technology output, but from our failure to invest effectively in innovation.
Concerns about a global productivity slowdown are rapidly spreading, as it is increasingly identified as one of the possible causes of the mediocre global growth performance. In a series of reports by The Conference Board we discuss the issues
(only for members) and have provided the accompanying measures
(publicly available) in detail. A recent report by the OECD
is addressing the topic as well.
The slowdown globally has indeed been dramatic, although not unprecedented. It seems we are, in some ways, reliving the dismal productivity story of the 1980s, an era during which many countries, including the United States, were struggling to find a path towards faster productivity and output growth – something they then accomplished in a spectacular way during the 1990s and early 2000s.
Trend growth of labor productivity (output per person), 1971-2014
Note: Trend growth rates are obtained using HP filter, assuming a l=100.
Source: The Conference Board Total Economy DatabaseTM, May 2015
Today, the debate about slow productivity is gaining traction again, especially in the United States. Following a sharp productivity boost immediately after the 2008/2009 recession, the growth path has been sluggish to say the least. And, despite similarities with the 1980s, the productivity collapse is extraordinary. In the United States, we have not seen such a slow labor productivity growth rate outside a recession for the past six decades.
The productivity slowdown is all the more surprising as there seems no shortage of new technologies. Many business leaders, policy makers, and analysts take it for granted that there must be a productivity boom right here and now as a result of these new technologies. The potential productivity gains from the combination of ubiquitous broadband and mobile, supported by cloud computing and big data analytics, are presumably huge, as reflected in the rise of apps economy and the sharing economy.
So if the macro productivity statistics don’t show it, mismeasurement of technology output could be a cause, as has been suggested in recent discourses in the media
, meaning that the productivity slowdown may not be real. Indeed there is an urgent need, just like in the 1990s, to improve the statistics on technology output. Inadequate price measures, a failure to measure consumer surplus (even though it is not supposed to be in the conventional GDP measure anyway) and, importantly, the inadequate reflection of the productivity gains from the apps economy in the output statistics, may all contribute to the mismeasurement story.
Recently, there has not been a comprehensive assessment of price change bias from new products and services for some time. There is an urgent need for the kind of measurement breakthrough we saw in the 1990s when price measures for PCs got adjusted for quality change, which drastically improved the measurement of productivity in the tech sector of the economy.
But despite the downward bias in today’s technology output how much will that matter for resolving the productivity puzzle? Probably not that much. History has proven over and over again that general purpose technologies – broadband and mobile probably qualify as such – do not automatically imply more productivity – immediately and big time. Even today, economic historians are still debating and analyzing how much of the growth during the industrial revolution can be allocated directly to the impact of the steam engine on growth in the 19th
century or the expansion of the electricity grid in the 20th
Without the benefit of perfect measures, one way to gauge the larger problem is to look at the pattern of productivity growth across industries. A convenient way to do is to construct so-called Harberger diagrams which plot the cumulative contribution of industry productivity growth to aggregate economy productivity growth against the cumulative share of these industries in aggregate value added.
There are two important observations to be drawn from this chart:
- Aggregate labor productivity growth dropped dramatically from 1.65 to 0.97 percent between 1999-2006 and 2007-2012.
- The curve is much less steep from 2007-2012 than from 1999-2006, indicating that the number of industries that are facing slower productivity growth has rapidly spread.
Indeed the value added of industries that showed negative labor productivity growth in the latter period increased from 14.4 percent (100-86.6) of aggregate value added from 1999-2006 to 35.1 percent (100-64.9) from 2007-2012. Some of the larger industries with negative labor productivity growth in the latter period are wholesale and retail trade and health care. Despite their measurement problems, tech-related industries like motion pictures and sound recording, compute & electronics products, and broadcasting and telecom are still among the industries with the fastest productivity increases. For an overview of the Harberger diagrams and individual industry contributions, see The Conference Board website
Cumulative Contribution of Labor Productivity Growth (Output per Hour) by Industry to Aggregate Productivity Growth in the United States
Source: Bureau of Labor Statistics and The Conference Board
It is hard to argue, on the basis of those metrics, that so many industries have recorded too little output and productivity growth, because technology output is in fact the key input for many of those industries. One important reason for the widespread slowdown is the fact that insufficient complementary investments have been made to make the new technologies really work.
For example, while many companies have spent fortunes on big data analytics and other ways to jump on the new technology bandwagon, they also need to invest in their organizational processes, complementary innovation, and in their people. Indeed today we should probably worry more about the lack of competencies across the workforce to make the new technology work than the failure to create new jobs with it. Buying more computing power alone is not a surefire way to increase productivity, nor is only hiring more data scientists.
The key factor to drive productivity is investment. In the 1990s it turned out that, once we measured the price changes for investments in computers and software well, we did find their impacts on productivity growth. Today, more than at that time, we need to focus on the complementary investment in intangible investments. Concerns about measurement might as well focus on those types of investments which are only partially measured as investment to begin with. The capitalization of R&D and software in the National Income and Product Accounts is an important step forward. But only once the other spending on intangibles, including workforce training, organizational innovations and marketing and branding, is also treated as investment rather than expenditure, we will get a sharper view of where the productivity gains from new technology have ended up.
Meanwhile, beyond improving the measurement, what are the other important things to do to tackle the productivity slowdown now? The new Conference Board report on productivity
focused on four important factors: (1) put the development of the workforce first; (2) invest in innovation; (3) let management practices lead productivity; (4) and allow competition to speed up resource reallocation. Differences in best practices between firms on those four aspects might well matter much more than the technology opportunities between different industries. With such an agenda, there is no reason for complacency. Blaming mismeasurement alone risks taking the eye off the ball of what more needs to be done.
The other reason for the widespread productivity slowdown is, no doubt, that most industries have also suffered from weak demand in the wake of the Great Recession. But as the productivity slowdown in most cases started well before the 2008-09 recession, the latter may not be the primary cause although it probably has made things worse.