Did innovation (and productivity growth) really slow down?

We live in a time of accelerating innovation, don't we? So why is it that economists can't measure it? Quite the opposite – productivity growth has been slow for decades.

Innovation is a kind of passepartout solution to all kinds of problems, whether big or small, systemic or not, long-term or current, pressing or of minor relevance. If something is wrong, we always at least try to innovate us out of trouble. But innovation is hard to measure. Economists refer to total-factor productivity (TFP), also known as multi-factor productivity, as a proxy for innovation.

In short, it’s a measure for economic efficiency – for example, more output with less (or, at least, the same) input. Hence, there is a direct link to economic growth: if TFP rises, ceteris paribus the gross domestic product (GDP) is supposed to grow as well. So when innovation happens, TFP increases, resulting in GDP growth. It is open to discussion if that’s truly the case, and we’ll briefly touch on this debate later.

For now, I’d like to focus on a seminal decline in TFP growth. As economists like Robert Gordon (The Rise and Fall of American Growth) have pointed out, TFP growth was fast from 1920 till 1970, and then again from 1994 till 2004. The rest of the time, it slowed down significantly. According to Gordon’s analysis, the decade from 1994 to 2004 reaped the fruits of the digital revolution. Everything that came after 2004 didn’t have any significant economic impact.

Let’s sink that in for a while. The last 15 years saw at least two major innovations: social media, with Facebook, Twitter et al., and the mobile revolution spurred by the iPhone and its Android counterpart. In economic terms, these innovations didn’t do much. For digital diehards like us, this is deeply disappointing and counterintuitive. And what about the stock market? Why are tech companies like Microsoft, Apple, Amazon or Google still the most valuable public companies in the world (not counting Saudi Aramco), and doing relatively well, even during the recent market crash?

Two possible explanations

If we look at the economic debate, there are roughly two possible explanations for this conundrum. The first one questions the measuring itself: is TFP growth really measuring innovation? And how is TFP itself measured, anyway? Is GDP growth a meaningful indicator of our collective wellbeing? Are we living in an era of a new economy (gasp) that can’t be measured with the old yardsticks?

Techno-optimists like Erik Brynjolfsson and Andrew McAfee (The Second Machine Age) belong to this camp. They firmly believe innovation is accelerating and the best is yet to come. The stock market, as always trading our expectations for the future, seems to agree. TFP growth might have been slow for the last 15 years, but this either will change in the future or results from a measuring error anyway.

The other camp accepts that TFP growth is slow as a fact and seeks for possible explanations. For starters, there is a huge base effect. The industrial revolution has propelled our Western economies to a level where further growth becomes very hard to achieve. Economist Dietrich Vollrath (Fully Grown: Why a Stagnant Economy is a Sign of Success) explains that slowing growth is the result of our own choices. In short: we now value other things, since the level of prosperity is simply high enough.

A second point is a simple nod to the thesis that technological progress is limited these days, compared to the breakthroughs of the industrial revolution, like electricity, running water, or the internal combustion engine. And then there is, at least in the US, the shift to a service-based economy: healthcare, education or personal care, to name a few, are fields where economies of scale are relatively weak and technological innovation is limited by default. Services comprise approximately two thirds of US consumption.

Productivity shifts to digital experiences

I’d like to add another point of view, and that is the shift to digital experiences. These typically are services with a strong technological foundation. Digital experiences show very strong economies of scale and are subject to technological innovation as well. However, so far, digital technology suffers from the productivity paradox coined by Robert Solow in 1987:

You can see the computer age everywhere but in the productivity statistics.

There’s of course a plethora of possible explanations for this paradox – but I won’t bore you with the details. Judging from history, it takes decades for technological innovation to manifest itself in economic terms like TFP and GDP. For example, the lightbulb, the internal combustion engine and wireless transmission were all invented in 1879. The economic impact of these major innovations appeared only over time.

The same may be true for information technology, the internet, smartphones and all the kinds of digital experiences that we are so excited about. The diffusion of technological innovation has clearly sped up since the middle of the 20th century. But more people adapting new technologies doesn’t equal higher productivity or economic growth. People, and companies, still have to figure out how to put these shiny new things to better use.

New companies have to emerge, and old companies must adapt or die. This takes time. And then, external factors like changing demographics might be way more important for TFP and GDP growth than technological innovation. (This is a point Dietrich Vollrath makes.) Perhaps it is still day one, the digital revolution has just begun, and we’ll see the true economic impact over the decades ahead. Only time will tell.