Issue Briefs

Humans will avoid the fate of the horses

Humans will avoid the fate of the horses

Martin Hutchinson

April 20, 2017

We are not horses. Yet the Luddite left is telling the story of how the horses lost their jobs after 1900 as a parable of what might happen to most humans as the robots take over all the jobs. In a truly free market, in which government was small and with little power, this would be impossible. The world economy is set up to benefit humanity, not horses, so the market would adjust to robots as a useful new tool to serve humans. Yet we are far enough from a free market today, and government is powerful enough and stupid enough, that the equine nightmare cannot entirely be ruled out. We must thus take steps to avoid being sent to the glue factory.

An emotional tale

Like the original machine-breaking Ned Ludd and his followers, the equine theorists have a nice emotional tale, with a little simplistic economic theory to back it. It is true that horses had lost market share in the human economy during the nineteenth century, with many of their transportation functions progressively replaced by railroads.

However, by 1900 this technological change had not reduced significantly the number of horses employed in the U.S. economy, which continued to increase through 1910 and beyond. It was the advent of the Ford Model T and equivalently low-priced, reliable tractors which put the horses out of a job. Their number began to decline, with the decline accelerating through 1960, by which time the horse population of the United States was only about 12% of its peak.

Workers are not horses

To the economically literate, there is just one flaw in this sad saga’s analogy to the position of low-skill workers now. At no time were the horses in a position of autonomous economic decision-making. They were pure consumers of resources, and in the vast majority of cases relied entirely on another species –humans– to provide them with the resources to consume. Had the humans disappeared, most of the horses would have starved in pretty short order.

Their position is not equivalent to that of truck drivers and assembly line workers who, provided they keep off the crystal meth, continue to be autonomous beings making economic decisions for themselves. What’s more, since humans make the consumption decisions in the market, other humans have a natural advantage over robots, horses, intelligent lizards or anything else as providers of goods and especially services to them.

More robots

One can imagine a completely free market with a limited labor supply in which robots would integrate seamlessly into the workforce. Indeed, from reports, this is what appears to be happening in Japan. The Henn-na Hotel in Sasebo, Japan (hat-tip to Andrew Stuttaford in National Review) has gone all-in on robotics, using them to clean the rooms, carry luggage and prepare food, with human staff used only for functions where guests benefit from interactions.

As usual with Japanese innovations, this is a signal of where we must eventually go. Japanese society has a shortage of unskilled labor to do physical tasks, because of its inverted pyramid age distribution (which leaves it short of broad, non-arthritic shoulders) and its admirable refusal to accept large-scale immigration. It has thus taken the lead in robot adoption, especially in the areas of elder care and personal services, where numerous unpleasant jobs currently exist, for which the cheap-labor crowd wishes to import Third Worlders.

Most common jobs

A recent analysis by Planet Money, using Census Bureau statistics, shows the most common jobs in each U.S. state, from 1978 to 2014. This is in certain respects dispiriting. For example the most common job in the District of Columbia has changed from the admirable but alas now uncommon “secretary” through the humble “janitor”, to the thoroughly economically damaging “lawyer.”

If the Trump administration cannot restore secretaries to their once-proud position in DC, let it at least restore the supremacy of the humble janitor; the rest of the country will be much better off!

Outside DC, the most common occupation in most states is “truck driver”, which predominates in 27 states, compared with only nine in 1978. This is the most horse-like of occupations, and one of those most likely to be eliminated by robotics.

However, even that elimination will take a generation; robot-driven trucks are still several years away from being available on a wide scale and their adoption will not then be universal immediately, because they will lack the “last mile” delivery capability of human-guided trucks.

In the 23 states in which “truck driver” is not the leading occupation, that occupation seems reassuringly non-robotic. Five lucky states still have “secretaries” leading the way, five have “primary school teachers” – surely an occupation pretty well impervious to robotics for decades yet – four have “computer software developers”, two nice old-fashioned places have “farm managers” and the rest are dotted around other service sector occupations, with New York having “nursing aides, orderlies and attendants.” Whereas a smattering of states in 1978 led with solid factory jobs that President Trump would like to bring back, in 2014 those jobs had already mostly gone and were the leading type of employment in no states.

Most traditional jobs not easily replaced by robots

Given all of the above, in about half the states, the leading occupation does not appear vulnerable to the robot revolution. Indeed, New York’s “nursing aides” may well appear as the leading occupation in a lot more states as the Baby Boomers age into senility.

Whereas Japanese seniors appear to accept robots happily as carers, it’s likely that U.S. seniors will be a lot more recalcitrant. Thus, there are many possibilities for the low-skill workforce (and for lawyers, gleefully predicted to become redundant by a number of observers) to retrain into new jobs as the robots arrive.

Regulators will complicate the issue

That is what is likely to happen in a free market. However, with the gigantic state, universal lobbyists and our infinite appetite to regulate everything, we cannot be sure that the benign free market result will occur.

For one thing, as former Labor Secretary candidate Andy Puzder has said, a higher minimum wage should be called “The Robot Employment Act” and he’s right. Environmental legislation also stunts the economy and restricts opportunities for modest-skill but high-paying jobs, in fracking for example. Conversely, the cheap-labor lobby seeking to fill the country with low-skill low-wage immigrants only makes the skills mismatch problem worse, in the U.S. at any rate.

There will be jobs for low skill people

Without those artificial stimuli to unemployment, low-wage workers should be able to compete with robots – they are after all serving food, they give haircuts or provide other services to human beings, not to robots. Therefore, they should have an advantage in that competition.

Indeed, even in the service sector, the best job opportunities for the modestly skilled will be working with robots to improve the overall quality of the services. In hotels, for example, robots will do the heavy lifting while trained humans will provide human faces with which guests can interact.

New training needed

Much training will be required. In that respect, the last two decades of the funny-money economy, in which employers have outsourced training to community colleges, have provided yet another artificial rip-off of the modestly skilled. The state should not be subsidizing cheapskate employers in this way. Working with robots will require new skills for employees; the employers buying the robots need to invest also in new skills for their workforce. Higher interest rates and a lower level of speculative nonsense, forcing companies to re-examine their operations on a long-term basis, are an urgent need.

In a free market, with humans continuing to control the system’s purchasing decisions, humans will be able to adapt to robots. Over the longer term, however, as robots take over an increasing share of the jobs, that adaptation may lead humanity as a whole to undergo two changes, neither of which should be resisted by governments.

Demographic adaptations

First, if robots can carry out much of the physical labor for which young, poorly educated men and women were previously used, the world probably needs fewer young, poorly educated men and women, and a lower population overall, since that will increase the resources per capita to give all the remaining humans a highly satisfactory living standard. The Japanese demographic with the inverted pyramid age structure is the ideal, with the hope that the global population will gradually shrink to its pre-industrial level of around 1 billion. That will enable our descendants in 2150 to live like Henry Fielding’s Squire Western.

Too many poor young people in emerging countries

The pyramidal demographic profile in many poor countries, with far more young people than older ones and population rapidly increasing, is thus thoroughly suboptimal, as indeed it is for poor countries’ attempts to become richer. Heaven forbid the world should attempt to change this by international coercion, but poor countries’ leaders are not stupid; they will quickly realize the benefits of domestic policies that produce an inverted-pyramid demographic result.

Genetically modified humans?

Second, as Yuval Harari has written, we will shortly have new techniques by which, either before birth or during their lifetime, humans can be “upgraded”, either genetically or by implanting chips that interact with the brain. Governments will attempt to resist this, on the grounds of its obvious inequality-producing potential (even though the costs of such techniques will drop rapidly once they are introduced.)

Nevertheless, assuming that this capability will be fully developed over time, if it will be combined with population restraint, it should allow the robot revolution to benefit everybody – those of limited intelligence will have IQ uplifts, while lawyers and other high-skill people made obsolete by robots will be re-programmed with more useful skills.

Do not restrict genetic engineering

Laws that restrict genetic engineering and implants are as futile and damaging as the pre-1896 British law that forbade automobiles except with a red-flag man walking in front. Like minimum wage legislation and artificially stimulated immigration, they could potentially turn us all into redundant horses.

There will be jobs

Low-skill people will continue to have job opportunities as robots take over, just different ones. High-skill people will also have roles to play, designing robots and software to make them work better with low skill people and later designing technologies whereby the capabilities of low-skill people can be enhanced. The only requirement is for meddling governments and their lobbyist hangers-on to stay out of the way.


Martin Hutchinson

Martin Hutchinson is a GPI Fellow and was a merchant banker with more than 25 years’ experience before moving into financial journalism. Since October 2000 he has been writing “The Bear’s Lair,” a weekly financial and economic column. He earned his undergraduate degree in mathematics from Trinity College, Cambridge, and an MBA from Harvard Business School.

This article was originally published on the True Blue Will Never Stain


The views and opinions expressed in this issue brief are those of the authors and do not necessarily reflect the policy of GPI.