‘Fools, Fools!’ shouted Benjamin, prancing around them and stamping the earth with his small hoofs. ‘Fools! Do you not see what is written on the side of the van?’
That gave the animals pause, and there was a hush. Muriel began to spell out the words. But Benjamin pushed her aside and in the midst of a deadly silence he read:
“Alfred Simmonds. Horse Slaughterer and Glue Boiler, Willingdon. Dealer in Hides and Bone Meal. Kennels supplied.” Do you not understand what that means? They are taking Boxer to the knacker’s!’
George Orwell, Animal Farm
But why did Napoleon, in George Orwell’s allegory of Animal Farm, send the horse Boxer to be killed and made into glue and dog food?
Napoleon was, of course, based on Joseph Stalin, and, being an unsentimental creature, he made the perfectly logical economic calculation that a weakened Boxer had come to cost more in feed than could be earned from his labour; ironically, this decision was true capitalism ‘red in tooth and claw’.
So how does this relate to the impact of technology on jobs and wages? Well, orthodox economists generally push back against the job destruction I outlined in Hiding from the Computers Part 1 by a) waving the theory of comparative advantage and b) looking at the historical wage record for unskilled workers since the industrial revolution.
For example, in an article in The New York Times entitled “Computers Jump to Head of the Class” earlier this week we see this:
Kazumasa Oguro, professor of economics at Hosei University in Tokyo, argues that smart machines should increase employment. “Most economists believe in the principle of comparative advantage,” he said. “Smart machines would help create 20 percent new white-collar jobs because they expand the economy. That’s comparative advantage.”
As an aside, the theory of comparative advantage is generally attributed to the English economist David Ricardo, but he was rather ambivalent about its impact on unskilled labour. He struggled with what he called ‘the machinery question’ and fluctuated from seeing technological advances benefiting all of society (divided into capitalists, land owners and workers) to a position whereby such advances could be detrimental to the working class. See, for example, here.
Comparative advantage is not as intuitive as absolute advantage. It is easy to see that if Napoleon is good at providing governance and security, and Boxer hard manual labour, then they should trade. Each has an absolute advantage in one area. The pig Napoleon is not as strong as the horse Boxer, but he is cleverer. However, as Boxer ages and becomes enfeebled, Napoleon becomes both cleverer and stronger than Boxer. Yet, at first, they should still keep trading despite Boxer’s weakened state. This is because Napoleon is still better at ruling than he is at working manually. Although he may be able to do manual work far better than the aged Boxer, this would require the opportunity cost of him giving up time ruling (and exploiting) the other animals on the farm.
Unfortunately for Boxer, there is a limit to this comparative advantage. Boxer requires a minimum calorific input to survive. As Boxer progresses toward senility, he saves Napoleon less and less time. At some point, a full day’s work by Boxer only saves, say, 10 minutes of Napoleon time. And that 10 minutes time spent ruling (exploiting) by Napoleon is worth less than Boxer’s feed—at that point, sorry Boxer: you are glue.
Too abstract for you? Well actually this is exactly what happened to Britain’s population of working horses early in the 20th century. The peak in horse numbers took place at 3.25 million in 1901 long after the industrial gathered pace. The economic historian Gregory Clark in his excellent book “A Farewell to Arms” explains what happened next:
But the arrival of the internal combustion engine in the late nineteenth century rapidly displaced these workers, so that by 1924 there were fewer than two million. There was always a wage at which all these horses could have remained employed. But that wage was so low that it did not pay for their feed.
By the end of WW2, UK working horse numbers had fallen further to a few hundred thousand. Technology had so radically changed the terms of trade between man and horse that around 90% of them got wiped out (with many of them boiled down and made into glue).
Of course, we still trade with horses. It’s a barter economy in which we feed them in return for them providing us with amenity value in allowing us to ride them. But only a few of them could occupy this niche; we just don’t need so many of them any more. (Incidentally, we also trade with cats and dogs: providing food, warmth and security in exchange for companionship and loyalty—although cats occasionally wander off to different homes.)
The key counterexample is animals. They have well-defined preferences, downward-sloping demand curves, demand for multiple commodities (if only both food and water, plus of course sex), and differential abilities. Yet most animals don’t trade with any other animals.
Why not? There are high fixed costs to trading at all. Most animals can’t overcome those costs. They aren’t smart enough. Lack of opposable thumbs, or lack of extended long-term trust, are other obstacles, not to mention “fear of being eaten.” There is nothing in a Walrasian model to rule out q = 0 no matter what kind of critters are walking around.
The point I am making here is that the theory of comparative advantage is not a religion. It is, however, a good theoretical explanation of how the division of labour together with trade can have a large welfare-enhancing effect. But it requires certain conditions to be in place. If they aren’t, comparative advantage doesn’t work.
Going back to our horses example, not only could they not price themselves sufficiently cheaply to compete with the man-tractor combination, but they also couldn’t exploit the new technology themselves.
Theoretically, it would have been much more welfare-enhancing for humans if at least some of the horses could have driven the tractors themselves. But, without opposable thumbs, they lack dexterity and are also poor performers in terms of manual dexterity and the ability to work in cramped spaces like tractor cabs (they are also, of course, a bit thick). I am now back full circle to the top of the table of technological bottlenecks from my previous post (click for larger image):
Humans, even unskilled workers, were able to surf from one technological innovation to the next during the industrial revolution because they possessed just those attributes that horses lack. And so we come to orthodox economist Exhibit 2 in humanity’s defence against the machines, the documented rise in real wages of unskilled labour throughout the industrial revolution. This chart is from Gregory Clark’s “Farewell to Arms” again:
The upward curve is a pretty stunning affirmation of both industrialisation and capitalism. Marx’s workers didn’t throw off their chains because they were too busy collecting their bulging pay checks—at least up until recently.
While I can see why Clark likes this chart—how often in economics do you find a time series going back 800 years—there are a few problems. The chart is of one occupation, in one country and it ends in 2000. We really want something that encompasses all unskilled labour and is more up-to-date. Further, I am switching to the US since a) it is the largest economy in the world, b) it is an economy where workers ability to resist the introduction of technology is minimal. To start with, let’s look at median wages using Census Bureau data (click for larger image):
Since the late 1970s, ‘all races’ is showing a minimum improvement despite a substantial rise in US per capita GDP. To pull out unskilled workers from these numbers, let’s look at a breakdown of wages by education level (from Acemoglu and Autor, click for larger image):
And from the Federal Reserve Board of St Louis database, the US labour force participation rate:
It appears, therefore, that unskilled labour can no longer pair up with new technologies and thus participate in the benefits of economic growth. They are either working for less or not working at all. This is a structural break from our experience gained from the first 200 years of the industrial revolution.
Nonetheless, many observers stress that we shouldn’t lose sight of the fact that technology will allow us to consume both more stuff and more services (I am putting climate change and resource depletion to one side for the minute). That is the rather Panglossian argument of Erik Brynjolfsson and Andrew McAfee in their very influential book “Race Against the Machine“. The Great Lisbon Earthquake of 1755 flattened a city, but Dr Pangloss still viewed this as “all for the best in the best of possible worlds”. Brynjolfsson and McAfee chronicle an epic technological upheaval that is smashing the post-war labour market to pieces, yet they end their book with this paragraph:
The third industrial revolution, which is unfolding now, is fuelled by computers and networks. Like both of the previous ones, it will take decades to fully play out. And like each of the first two, it will lead to sharp changes in the path of human development and history. The twists and disruptions will not always be easy to navigate. But we are confident that most of these changes will be beneficial ones, and that we and our world will prosper on the digital frontier.
That sounds optimistic. Yet it doesn’t really tell us how the unskilled will escape the glue factory. I will return to this in my next post.