I feel that productivity is one of those elusive ideas. So appealing in principle that you decide to put them to work, only to discover that what you get is not exactly what you saw.
Productivity is a recurring theme in this blog. How couldn’t it be? The ultimate goal of technology and innovation is to increase the quantity and quality of things we produce and enjoy. The ability to make more stuff with less work is what allows us to enjoy more things which, like smartphones, would have been unimaginable luxuries—magic—only a few decades ago. The problem is that it is not easy to measure neither the quantity nor the quality of the things we produce and consume. How many facebookies do you eat every day? How much better has been your last report after the last tweaks to your favourite search engine?
And here comes the debate. Despite ongoing IT-related innovation, aggregate U.S. productivity growth slowed markedly after 2004. While economists are again unable to find the productivity in their statistics, many people in Silicon Valley think that this slowdown has to be at least in part illusory. They argue that information technology is providing significant benefits that just don’t show up in the standard measurements of wages and productivity. The “productivity paradox 2.0” is alive and kicking.
I am with Tyler Cohen when he says that “while many Internet entrepreneurs are economic heroes, statisticians are also pretty good at what they do.” And it seems that statisticians has been putting recently a bit of hard work to conclude that the productivity slowdown is all too real. Two papers recently published have brought the productivity debate back to the forefront.
In “Challenges to Mismeasurement Explanations for the U.S. Productivity Slowdown,”(1) Chad Svyerson shows how a number of disparate analyses offer empirical challenges to the “mismeasurement hypothesis”. The productivity slowdown has occurred in dozens of countries, and its size is unrelated to measures of the countries’ consumption or production intensities of information and communication technologies. The productivity slowdown is too big in scale, relative to the size of the tech sector, to be plausibly compensated for by tech progress. Estimates of the surplus created by internet enabled digital technologies fall far short of the $2.7 trillion or more of “missing output” resulting from the productivity growth slowdown.
In “Does the United States have a productivity slowdown or a measurement problem?”(2) David M. Byrne, John G. Fernald, and Marshall B. Reinsdorf, report that they also find little evidence that the slowdown arises from growing mismeasurement of the gains from innovation in IT-related goods and services. Mismeasurement of IT hardware is significant prior to the slowdown. Furthermore, if we fully accounted for the impact of technology, the recent slowdown in labour productivity would look even worse than it does in the official statistics.
The decline in economic dynamism—both in the form of fewer start-ups and slower reallocation of labour resources in response to productivity shocks—supports the idea that productivity-enhancing innovations are diffusing through the economy more slowly
If not mismeasurement, why did productivity growth slow? A plausible story is that the fast-growth 1995-2004 period was the anomaly—a one-time upward shift in the level of productivity rather than a permanent increase in its growth rate. The mid-2000s slowdown should be seen as a “return to normal”—marking an end (or pause) in a phase of exceptional, broad-based gains from the production and use of information technology.
Okay, but end or pause? How do information and communication technologies stack up with past major general-purpose technologies?
The maximum impact of steam power on British productivity, the benefits of railroads, or the general availability of electricity in the US took decades to have a real impact. The internet can be seen as a major follow-up invention from the microchip, just as the internal combustion engine led to the interstate highway system in the United States 60 years later.
In a previous paper(3), Svyerson shows that labour productivity increased on average by 2.4 percent annually after the inventions of the internal combustion engine, electricity, and running water between 1891 and 1972. Then it slowed, averaging only 1.4 percent annually between 1973 and 1996. Between 1996 and 2004, it surged again, growing 2.5 percent annually, which scholars ascribe to the fast spread of the internet.
Looking forward, we could get another wave of the IT revolution. Indeed, it is difficult to say with certainty what gains may yet come from cloud computing, the internet of things or even a radical increase in mobility enabled by smartphones.
On the other hand, Robert Gordon is convinced that IT revolution does not have the same lasting impact as previous industrial revolutions. Gordon is not particularly impressed by Facebook or the iPhone. He comes back now with new book, “The Rise and Fall of American Growth,” where he argues that what our economic statistics hugely undercount, is the welfare gains of the 19th and 20th centuries.
GDP didn’t capture how much better it was to walk on streets that weren’t covered in horse manure or how much more pleasant the summer is with air conditioning. Compared to the invention of indoor plumbing, electricity and the internal combustion engine,
In other words, we have no idea of what’s going to come.
For the time being, and now that Moore’s Law seems to have reached an end, I am afraid economists and their dismal statistics will end up winning the battle. This week, The Economist is surprisingly optimistic and suggests that software can make up for the loss of exponential improvements in hardware, which I’d very much doubt were it not for the outrageous defeat of Lee Se-dol.
(1) Chad Syverson (2016). Challenges to Mismeasurement Explanations for the U.S. Productivity Slowdown NBER Working Papers DOI: 10.3386/w21974
(2) Byrne, D., Fernald, J., & Reinsdorf, M. (2016). Does the United States have a Productivity Slowdown or a Measurement Problem Finance and Economics Discussion Series, 2016 (017), 1-74 DOI: 10.17016/FEDS.2016.017
(3) Chad Syverson (2013). Will History Repeat Itself? Comments on “Is the Information Technology Revolution Over?” International Productivity Monitor, 25, 37-40