In writing frequently about American labor force productivity, and arguing vigorously that it can and should become the critical focus of politics and policymaking in the near future, I inevitably encounter the forceful pushback that productivity is “mismeasured” due to “technology.” The critics’ implication is that American productivity is actually higher than it is reported, and if productivity is understated, then we shouldn’t pay attention to it, nor attempt to intelligently increase it. Unpacking this sentiment is delicate because the skepticism over productivity is rooted in some genuine economics that are quite arcane, but which have then been corrupted by the mainstream narrative into something that really is no longer true.
First, let’s look at the genuine economics. Productivity is real economic output per hour, and for the economy as a whole, that’s real GDP / hours worked. This seems simple enough, since we have widely understood statistics for both GDP and hours worked. The challenge originates in the word “real.” Because real GDP stats are adjusted for inflation, the methodology for adjusting for the falling price of information technology impacts our reported GDP, and by extension, productivity. In general, if a given item of technology, say a computer or a robot, costs the same as it did 10 years ago in nominal terms, but is many times more powerful, then statistically that piece of technology has undergone price deflation. The domestic production of these physical items whose prices are statistically declining (but not actually declining) increases the reported real GDP, and by extension, increases reported productivity all else equal. This sounds like an overstatement, not an understatement, right? Well the effect that economists note is that in recent years the rate of technology price declines baked into real GDP and productivity stats has been less negative, meaning that in the post-crisis period, the statistical boost from deflating tech prices has been less positive.* This effect has led to the suspicion that perhaps our productivity has been less bad than the statistics indicated over the last 6 years or so. My personal view is that the higher pre-crisis numbers were more distorted by this particular effect, but that’s less important. The issue is that this intricate quirk, which might suggest we have understated the value of technology production, has morphed into the following belief on technology and productivity (which I believe to be a fallacy):
The idea I hear most during lamentations on “mismeasurement,” is the feeling that our productivity must have been enhanced by the existence of time-saving and money-saving technologies like Google, Facebook, and Twitter, and because these things are free for consumers, that somehow their impact has not been captured. In my view, the impact of these technologies on the productivity of the American workforce is captured by the data, and in some cases it may not be positive. Let’s look at an example. Google Maps and competing services give away real-time traffic and navigation services for free, which has eliminated the expense and time associated in the past with procuring paper maps. For map makers, this development is catastrophic, and their production drops, thereby reducing GDP and productivity (of their workers). This effect must be made up for, and more, by increased production per hour by the users of Google Maps. Time spent sitting in traffic or asking for directions or hunting down maps of obscure meeting locations must now be spent in genuine productive use. If Google Maps users are successful in raising their own output (specifically making extra money for themselves and/or their employers) during their working hours because they can work more efficiently using the technology, then measured GDP and productivity will rise. If however that increased efficiency merely results in lower stress / lower effort working hours, where the monetary value of the output remains unchanged, neither GDP nor productivity increases. While the hassle of procuring maps has been removed, if that effort is not redirected to something more (financially) productive, the stats don’t (and shouldn’t) move.
The productivity statistics do reflect the labor force’s ability to translate available efficiency gains into extra output per hour, measured financially. Importantly, they reflect the positive contribution of unicorn technologies which have revolutionized efficiency in certain tasks, as well as any negative contribution from leisure-focused apps which provide potential distractions during working hours.
So I’m not concerned about the statistics, but based on what might seem like a technicality if it feels inappropriate for the statistics to focus so intently on the monetary value of output. This begs another question: is a modern, more technologically-streamlined workday a more “productive” workday even if the financial output and hours worked don’t change? Not according to traditional economists. I’ll tackle this in a future blog.
*The reasons for this are complex, but substitution of imported technology in lieu of domestic production is part of it. Read more here: https://www.federalreserve.gov/econresdata/notes/feds-notes/2015/recent-slowdown-in-high-tech-equipment-price-declines-some-implications-for-business-investment-labor-productivity-20150326.html