There is a popular meme going around, popularised by the likes of Tyler Cowen, Paul Krugman and Noah Smith that suggests that recent falls in worker compensation as a percentage of GDP is mostly due to the so-called “rise of the robots”:
For most of modern history, two-thirds of the income of most rich nations has gone to pay salaries and wages for people who work, while one-third has gone to pay dividends, capital gains, interest, rent, etc. to the people who own capital. This two-thirds/one-third division was so stable that people began to believe it would last forever. But in the past ten years, something has changed. Labor’s share of income has steadily declined, falling by several percentage points since 2000. It now sits at around 60% or lower. The fall of labor income, and the rise of capital income, has contributed to America’s growing inequality.
In past times, technological change always augmented the abilities of human beings. A worker with a machine saw was much more productive than a worker with a hand saw. The fears of “Luddites,” who tried to prevent the spread of technology out of fear of losing their jobs, proved unfounded. But that was then, and this is now. Recent technological advances in the area of computers and automation have begun to do some higher cognitive tasks – think of robots building cars, stocking groceries, doing your taxes.
Once human cognition is replaced, what else have we got? For the ultimate extreme example, imagine a robot that costs $5 to manufacture and can do everything you do, only better. You would be as obsolete as a horse.
Now, humans will never be completely replaced, like horses were. Horses have no property rights or reproductive rights, nor the intelligence to enter into contracts. There will always be something for humans to do for money. But it is quite possible that workers’ share of what society produces will continue to go down and down, as our economy becomes more and more capital-intensive.
So, does the rise of the robots really explain the stagnation of wages?
This is the picture for American workers, representing wages and salaries as a percentage of GDP:
It is certainly true that wages have fallen as a percentage of economic activity (and that corporate profits as a percentage of economic activity have risen — a favourite topic of mine).
But there are two variables to wages as a percentage of GDP. Nominal wages have actually risen, and continued to rise on a moderately steep trajectory:
And average wages continue to climb nominally, too. What has actually happened to the wages-to-GDP ratio, is not that America’s wage bill has really fallen, but that wages have just not risen as fast as other sectors of GDP (rents, interest payments, capital gains, dividends, etc). It is not as if wages are collapsing as robots and automation (as well as other factors like job migration to the Far East) ravage the American workforce.
It is more accurate to say that there has been an outgrowth in economic activity that is not yielding wages beginning around the turn of the millennium, and coinciding with the new post-Gramm-Leach-Bliley landscape of mass financialisation and the derivatives and shadow banking megabubbles, as well the multi-trillion dollar military-industrial complex spending spree that coincided with the advent of the War on Terror. Perhaps, if we want to look at why the overwhelming majority of the new economic activity is not trickling down into wages, we should look less at robots, and more at the financial and regulatory landscape where Wall Street megabanks pay million-dollar fines for billion-dollar crimes? Perhaps we should look at a monetary policy that dumps new money solely into the financial sector and which has been shown empirically to enrich the richest few far faster than everyone else?
But let’s focus specifically on jobs. The problem with the view that this is mostly a technology shock is summed up beautifully in this tweet I received from Saifedean Ammous:
The Luddite notion that technology might render humans obsolete is as old as the wheel. And again and again, humans have found new ways to employ themselves in spite of the new technology making old professions obsolete. Agriculture was once the overwhelming mainstay of US employment. It is no more:
This did not lead to a permanent depression and permanent and massive unemployment. True, it led to a difficult transition period, the Great Depression in the 1930s (similar in many ways, as Joe Stiglitz has pointed out, to the present day). But eventually (after a long and difficult depression) humans retrained and re-employed themselves in new avenues.
It is certainly possible that we are in a similar transition period today — manufacturing has largely been shipped overseas, and service jobs are being eliminated by improvements in efficiency and greater automation. Indeed, it may prove to be an even more difficult transition than that of the 1930s. Employment remains far below its pre-crisis peak:
But that doesn’t mean that human beings (and their labour) are being rendered obsolete — they just need to find new employment niches in the economic landscape. As an early example, millions of people have begun to make a living online — creating content, writing code, building platforms, endorsing and advertising products, etc. As the information universe continues to grow and develop, such employment and business opportunities will probably continue to flower — just as new work opportunities (thankfully) replaced mass agriculture. Humans still have a vast array of useful attributes that cannot be automated — creativity, lateral thinking & innovation, interpersonal communication, opinions, emotions, and so on. Noah Smith’s example of a robot that “can do everything you can do” won’t exist in the foreseeable future (let alone at a cost of $5) — and any society that could master the level of technology necessary to produce such a thing would probably not need to work (at least in the sense we use the word today) at all. Until then, luckily, finding new niches is something that humans have proven very, very good at.