An NBER paper “Robots are us – Some Economics of Human Replacement” paints a grim picture of our robotic future, in which the robots undermine their customer base, making the vast majority of humanity redundant. As good academic-institution social democrats, the paper’s authors Seth G. Benzell, Laurence J. Kotlikoff, Guillermo LaGarda, and Jeffrey D. Sachs then suggest the solution lies in ever greater redistribution schemes. Since I find their solution ideologically repugnant (and would loathe living in a world in which it had been implemented), I thought it worth examining their thesis, to determine whether, even if it is correct, there is a better way out of this ultimate human quandary.
The paper builds a model by considering robots as accumulated software, the output of “dead high-tech workers.” It then examines the fate of high-skill and low-skill workers as the population of robots increases and their sophistication grows. Over time, high-tech wages, after increasing, begin to decline and more and more high-tech workers move into the area of personal services, assumed to be those activities where humans cannot be replaced by robots.
With realistic assumptions about savings rates, wages decline, as does the capital stock, and over a generation workers are immiserated, with the youngest workers suffering the worst falls in their lifetime standard of living. The authors perform a number of sensitivity analyses, but find that the only trajectory that produces higher living standards is that with much higher savings rates, in which the stock of capital increases far enough to allow higher wages for the entire workforce. They also conclude, interestingly, that code should be non-rival but excludible, in other words that it should be a private good but with cheap and easy licensing to get round patents.
The authors draw four conclusions. First, they expect a long-run decline in labor’s share of income. Second, they expect a highly cyclical robotized economy, with Kondratieff-like long waves. Third, they expect current output to depend increasingly on pure software investment, so that Silicon Valley will rule us all. Finally, they recommend more vigorous redistribution, without suggesting how that will restore wage rates, but simply to equalize the misery.
The first conclusion we can draw from the study is a macroeconomic one. If higher savings rates would alleviate the problem of human immiseration through replacement by robots, then we must take steps to raise savings rates, a problem in the United States for the last two decades. As readers of this column will know, the best way to achieve this is to increase interest rates, pushing then substantially above inflation, ending the two decades (as of this month) during which rates have been kept artificially low by the Fed. The U.S. economy has already been substantially decapitalized through Fed policy, while outsourcing to cheaper wage areas has been encouraged by artificially low capital cost differentials between the U.S. and emerging markets. This alone is responsible for much of the decline in labor’s share of the U.S. economy that is noted and deplored by the authors.
The authors’ principal structural solution, more redistribution, would merely share the misery of lower living standards. It’s a wish shared generally on the left, where a “Minimum Living Wage” movement has gained considerable traction. This is especially foolish; if demand for labor has been reduced to unacceptable levels by robotization, then higher minimum wages, imposed on employers, will simply reduce the demand for labor further. If McDonalds workers must be paid $20 an hour, and robots are universally available and capable, then guess how long it will take to robotize McDonalds and put all those workers out of a job?
Both the authors and previous robo-pessimists back to Maynard Keynes in 1930 have suggested that the real problem caused by robotization would be one of insufficient work at any wage, requiring a mass extension of welfare provision, taxing the remaining few productive workers to provide subsistence for the unemployed masses, perhaps even a majority of the population – Silicon Valley calls this the “Universal Basic Income.” This is a true dystopia; with half or more of the population existing on welfare and having no purpose in life, the consumption of illegal substances would soar, as would radical movements and criminal activity.
The idea that the great mass of the population could be deployed in artistic activity is nonsense; the great majority of people have no significant talent for it, nor sufficient interest in it to make artistic activity the focus of their life. As my grandmother used to say, with absolute conviction, quoting Dr. Isaac Watts’ 1715 masterpiece “Divine Songs Attempted in Easy Language for the Use of Children:” Satan Finds Some Mischief Still For Idle Hands To Do.
With half the population idle and relatively impoverished after a decline in living standards, however generous the redistribution, the “mischief” level would be intolerable. If world population levels were much lower than today, and the resources available per individual correspondingly higher, the unemployed could doubtless be occupied in expensive but enjoyable activities, but absent a major war, that development appears at least 200 years off.
The twin problems, of high minimum wages putting everybody out of work and high welfare benefits turning them into feckless criminals, can perhaps be mitigated in the U.S. by expanding the Earned Income Tax Credit, which rewards work without increasing the costs of employment. However, since of all tax programs this is said to be the most subject to fraud, even this solution seems unlikely to be effective.
The political system’s natural response to the threat of robotic redundancy will be to make “job-destroying” robotization illegal. We can already see this at work with the Obama administration’s rule for the use of drones, which requires them to be flown only in the line of sight of the flyer, an obviously unworkable rule if any of the benefits of private sector drone use are to be realized. Similarly, it’s likely there will be attempts to prevent the adoption of self-driving cars and trucks, a development which might put truck drivers out of business but would be hugely beneficial to the economy as a whole. With those approaches, the pre-1896 rule that automobiles required a man in front with a red flag would have been maintained, and horse-drawn transportation would have been mandated on the grounds that otherwise there would be no employment for the unfortunate horses.
There are two technological solutions that are likely to eliminate the problem of robot-driven redundancies, provided regulators do not get in the way. One is the new field of “brain-computer interfaces” by which computers and human brains are able to interact directly, with brains sending signals that can be interpreted by clever software, and vice versa. There are currently a number of small enterprises producing early versions of these, without a great deal of corporate structure or venture capital funding; it is a little like the genesis of the PC industry in the mid-1970s.
As yet, devices have not been produced that can be easily monetized, and the market is tiny, but it seems likely that as more useful and effective brain-computer interface mechanisms are developed, the devices will come into general use, revolutionizing human capabilities. Given the nature of the products concerned, it is even possible that they will emerge before robotization has gone much further – the development of useable robots has after all been remarkably slow, compared to other developments in the tech sector. In any case, human brains equipped with direct sophisticated links into computers will be much more capable than humans alone – and will correspondingly be able to undertake much higher-level jobs, many of which are doubtless as yet unimaginable.
The other potential advance, even more prone to subversion by regulators, is direct genetic manipulation to improve the intellectual capabilities of mankind. Even in our current state, intelligent people are more capable of amusing themselves non-destructively than stupid ones. Should human capabilities be genetically enhanced significantly, then it likely that some people would be able to find new unimaginable sources of employment producing new unimaginable products and services, while even the lesser intellects would be able to enjoy Keynes’ 1930 dream of a 15-hour workweek followed by ample leisure enjoying the artistic, musical and creative output of mankind. If each person had sufficient intellectual resources, Satan would be thwarted and mass leisure would be no bad thing.
As usual, the free market has potential solutions to the problem of robot redundancy, if it is allowed to reach them. By improving human capabilities through machines that help humans to function, and by improving humans themselves through genetic engineering, we can ensure that human development keeps up with robotics, so that the human race’s potential is maximized, and robots become invaluable helpers in achieving that potential. By this means a long-term utopia is attainable, in which a more limited population of humans can lead lives both productive and leisured at very high standards of living, with their Downton Abbeys staffed by impeccably behaved robot servants.
“Robots are us” is a useful dystopia, showing us the nightmare of a society in which dead computer programmers embodied in robot software combine to reduce the life potentials of the living. Given the misguided tendencies of the world’s policymakers, it is even a likely dystopia. But it is not an inevitable one.
(The Bear’s Lair is a weekly column that is intended to appear each Monday, an appropriately gloomy day of the week. Its rationale is that the proportion of “sell” recommendations put out by Wall Street houses remains far below that of “buy” recommendations. Accordingly, investors have an excess of positive information and very little negative information. The column thus takes the ursine view of life and the market, in the hope that it may be usefully different from what investors see elsewhere.)
Martin Hutchinson is the author of “Great Conservatives” (Academica Press, 2005) — details can be found on the Web site www.greatconservatives.com and co-author with Professor Kevin Dowd of “Alchemists of Loss” (Wiley – 2010). Both now available on Amazon.com, Great Conservatives only in a Kindle edition, Alchemists of Loss in both Kindle and print editions.