Everyone's worried. Even Jack Ma thinks that his job is in danger from robots. And he's not alone. Nary a planetary rotation completes without a warning about the dire consequences of the ongoing robot uprising. Everyone is vulnerable: truck drivers, cab drivers, and even CEOs.
At the danger of repeating Paul Krugman's mistake in 1998 of forecasting that the Internet's impact would be no more than the FAX machine, if the robots are taking over, why is productivity growth at generational lows?
It's an old argument, perhaps typified by the Luddites destroying machinery that was putting them out of work (and maybe not quite fairly; there is modern argumentation that the Luddites were taking collective and property-destroying action as a reaction to their employers exercising their improved bargaining power; I frankly see little distinction).
The real problem with the industrialising United Kingdom (and Russian serfdom), was not technology. It was the threat and materialisation of state-sponsored violence to maintain class control.
A few days ago, William Baumol died. He contributed an observation, which we call Baumol's Cost Disease, which found that real wages in low productivity industries actually rose like those in high productivity industries, as workers left low-productivity industries for higher productivity industries. This simultaneously increases their income, as well as forces the employers in the lower productivity industries to raise salaries to retain and attract replacement labour.
What Baumol called a dilemma, and what we have come to describe as a disease, is actually precisely the salve for technological productivity gains to be distributed, even to parts of the economy that aren't experiencing productivity growth. The real disease would be if Baumol's dilemma didn't exist — but it does.
Of course, the story gets more complicated in the decades after Baumol's initial observation. In the past seven decades, real hourly compensation is up roughly two-fold, while productivity is up nearly four-fold, with the bulk of this divergence occurring after 1980. Karl Marx would refer to this gap as surplus value — the product of labour which it does not get to keep. The classical misinterpretation is that the surplus value flowed to owners of capital. This is wrong. As we've previously observed, the bulk of this surplus value had flowed to consumers, and to consumers who are not producers, subsidised by transfer payments.
The reality is that liberal economics and technology liberated most of the world from slavery, serfdom, and state violence enforced class control in the 19th and 20th centuries. The second reality, which often conflicts with the ideological priors of the exponents of the first reality, but is no less true, is that the mixed economy has promoted Baumol's Salve (yes, I am leading the charge to rename this phenomenon).
Take California as an example, which is perhaps America's preeminent example of the mixed economy: it has outperformed US Real GDP by more than 15% cumulatively in the past two decades, despite experiencing nearly identical decline in manufacturing employment (and starting manufacturing labour intensity) as the entire country.
It becomes quickly apparent that the losers in automation aren't losing because of automation — they are losing because poor political management.
Productivity growth is capital substitution, and robots and AI are just another form of capital substitution. This has occurred from the beginning of time. Further, since the Great Recession ended, this trend has been slowing, not accelerating. And it has not eaten labour — Real Median Wages are at their recorded historical highs.
Capital substitution is expensive, and the labour pool that can facilitate it is very small. Perhaps 0.5% of the labour force are programmers, and most of them are working to maintain and fix existing systems. And, while everyone has spent their cycles worrying about capital substituting manufacturing labour, employment in manufacturing has risen by around a million payrolls since early 2010:
The real trend in technology in the past decade hasn't been to substitute labour in production. That was the trend in the prior decades. In fact, labour-intensity of manufacturing has not changed in seven years, after falling nearly every year for the prior four decades.
Productivity growth is the normal state of affairs, but the acceleration of manufacturing productivity coincided with the steep economy-wide productivity that peaked in the late 90s/early 2000s.
But, as you can plainly see, that trend quickly reversed itself, and we are now experiencing the weakest manufacturing productivity growth (and highest change in labour-intensity) in series history.
This all isn't to say that technology-driven productivity growth won't accelerate once again. It almost must. Global decline in birth rates decades ago set low labour force growth in motion, and capitalists will do their best to compensate. But hearing the cries about job-stealing robots sounds a great deal to me like early 20th century sci-fi describing how 2017 might look. We do know that it will be different, with much whacky and wonderful new technology, but the nature of scientific discovery is that it's rarely exactly what we looked for, and never what we expected. The domain which the robots have taken over is that of our imaginations.
Finally, it also turned out that Paul Krugman wasn't that all that wrong. The crux of his argument was about the technology boom's impact on productivity (and not a commentary on how different our world is in the age of selfie sticks and dick pics):
Productivity will drop sharply this year. Nineteen ninety-seven, which was a very good year for worker productivity, has led many pundits to conclude that the great technology-led boom has begun. They are wrong. Last year will prove to have been a blip, just like 1992.
I'm sure he regrets comparing the Internet to the FAX machine, but the towering productivity growth that existed at the time (see figure 1) has subsequently fallen to nearly the lowest ever from nearly the highest ever.
He nailed a couple of other things in the same piece:
Within two or three years, the current mood of American triumphalism — our belief that we have pulled economically and technologically ahead of the rest of the world — will evaporate. All it will take is a few technological setbacks or a mild recession here while Europe or Japan recovers a bit.
And this gem:
Sometime in the next 20 years, maybe sooner, there will be another '70s-style raw-material crunch: a disruption of oil supplies, a sharp run-up in agricultural prices, or both. And suddenly people will remember that we are still living in the material world and that natural resources matter.
Not bad, Krugman.