Should we Fear the March of the Robots?
Courtesy of Acting Man, Pater Tenebrarum
A Jumble of Memes
It has recently become quite fashionable to drag up the old Luddite argument that technological progress will destroy jobs, which is to say, destroy them on a net basis. Allegedly, the “rise of the Robots” will accomplish what centuries of economic and technological progress in the capitalist market economy have failed to do.
As you will see in the video below, the purveyors of this idea assure us that “this time, it's different”. Allegedly, humans are now in the same position that horses found themselves in when the automobile was invented. If you are not groaning inwardly by the time this argument is proposed, then you urgently need to brush up on your knowledge of economics (the same obviously goes for the makers of the video).
Closely associated with this idea is the meme of a “basic income for everyone”, to be provided by the State. People arguing in favor of such a program have apparently forgotten that nothing really comes for “free”. Not only must this redistribution be funded by others who produce real wealth and the State would have to forcibly expropriate them in order to provide this basic income, but we may rest assured that this “gift” will come with strings attached. A crypto-communist State providing basic incomes to all citizens will strip them not only of the incentive to work, it will inevitably also abridge their freedom in many other spheres of life. The State never provides something for “free”.
Another close cousin of this “free income” idea is a recent proposal mentioned in “Foreign Affairs”, that central banks should not only keep printing money, but should alter their modus operandi by paying this newly printed money directly into the accounts of all citizens, so as to spur demand. Not only is this proposal based on the economic fallacy that spending and “demand” are the causes of economic growth, it also fails on the grounds that the authors cannot pick and choose between the inflationary effects that will result. In other words, they won't only get the “good stuff” that they like, but also all the bad stuff that ultimately far outweighs the former.
This is beside the fact that printing an additional quantity of money cannot alter the amount of real capital available to society, and hence cannot possibly increase general prosperity. Of course the current modus operandi of central banks is not any better – it merely leads to a different wealth redistribution than the one that would result from this proposal. In practice, the current method largely tends to confine the inflationary effects to a shift in relative prices in the economy, and the spurring of bubbles in financial asset prices. It is important though to keep in mind that this is not guaranteed, as it depends on contingent data. There are also other, unseen effects, that must be considered. For instance, it is argued that “consumer prices are more or less stable” and hence no “bad” inflation is in sight. What those arguing along such lines forget is that absent the inflationary policy, consumer prices may actually be declining, which would greatly increase real incomes. What the precise effects of a distribution of newly printed money into the accounts of all citizens would be will depend on their individual decisions as to how to deploy that money (it will presumably still be open to them to decide whether to add to their cash balances, invest or consume).
Finally, we would note that this proposal shows once again how incredibly one-sided the whole mainstream debate over central bank policy has become. Apparently, no-one in the mainstream can think of a “solution” to our economic problems anymore that does not involve gobs of money printing. Almost every editorial on the topic contains some variant of an inflationary plan. Evidently, what will be required is a complete catastrophic collapse of the monetary system to shake these people out of their trance. The catastrophes of the past have evidently been forgotten.
Technological Progress and Employment
Let us get back to the alleged threat emanating from progress in robotics. First, here is the video mentioned above. It is entitled “Humans Need not Apply” and has gathered over 2.2 million views on You-tube already, testament to the enduring popularity this essentially Luddite meme.
Humans Need not Apply – the Robots are Coming for our Jobs!
The makers of the video state at the end that “well, there is anyway nothing that can be done about this, but we need to be prepared” – or in other words, “we” must do something, anyway. Since they are not telling us what should be done, we are left to infer what could be meant. Presumably, they don't want us to burn researchers working in robotics on the stake, or go out and smash robot factories.
That leaves only intervention by the State: the placing of restrictions on what robots may be used for, the payment of a “universal basic income” such as mentioned above, or other forms of interventionism. To this we can only say: we all should be grateful that during the all too brief age of classical Liberalism, those who sought to use the State to impose restrictions on the market economy were unsuccessful for several decades. If not for the capital accumulation this age made possible, we very likely would not need to worry about robots today, because the alleged problem would still be decades away.
Let us briefly consider a few empirical facts demonstrating the power of the market economy. A little over two centuries ago, nearly 90% of the population worked in agriculture. This percentage has declined to just above 2%, while output today is orders of magnitude greater:
A massive, inexorable decline in agricultural employment, which went hand in hand with a no less impressive increase in output – click to enlarge.
We couldn't find a chart showing the growth in agricultural output over the entire period, but since 1948 alone, US far output has increased about 2.5 times, amid soaring productivity:
Farm output since 1948 – productivity soars – click to enlarge.
A similar productivity revolution has been underway in manufacturing for quite some time, with the result that fewer and fewer workers produce an ever larger amount of goods:
Manufacturing jobs as a percentage of total employment vs. industrial production – click to enlarge.
As a result of all this, the character of employment has changed dramatically, as the next illustration shows:
Changes in labor force distribution since 1840 – click to enlarge.
We should perhaps also mention that instead of the often predicted “Malthusian catastrophe” theory, which posited that the world would eventually be unable to feed its growing population, the exact opposite has in fact happened: agricultural output has not merely kept up with global population growth, but there has been enormous growth in per capita food production as well.
Global corn, wheat and rice production vs. global population since 1961 – click to enlarge.
Evidently, a main object of economic activity is to produce more with less, i.e., it is aimed at alleviating scarcity, by using less and less input to produce more and more output. Not only that, the division of labor and the accumulation of capital have enabled an enormous lengthening of the capital structure and a vast increase in specialization, which allows us to produce goods that could otherwise not be produced at all.
People have been relieved of the tedious drudgery of times past (as illustrated by the US steel industry producing 56,000 scythes and 20,000 pitchforks per year in the 1830s). As a result, they enjoy far more leisure than before, even while their income and wealth have increased enormously. It should not be necessary to mention it, but people considered relatively poor in today's developed nations enjoy amenities that a king of yore could not even dream of.
Capital and Progress
However, in the video it is argues that none of this matters. Allegedly, “this time, it is different”: the increased use of robotics will make so much employment redundant, that the increase in production this will presumably enable will be wasted on a penniless army of unemployed people.
We are told that it would be foolish to expect that things will continue as they have up until now; not enough new jobs and job categories will be created to replace those that are lost. We are like horses, and the robots are our version of the automobiles that made horses superfluous. The makers of the video neglect to mention that horses did not build the automobiles that made them redundant – humans did.
In order to prove their case, the authors show us statistics of the number of people employed in job categories that they believe are likely to disappear in the future. They compare these numbers to the number of those employed in the computer industry, which they assert proves that the creation of new job categories cannot keep up with the jobs that are lost. One of the examples given are driver-less cars. What will all those truckers and other people now employed in transportation do once driver-less cars become ubiquitous? Can we really expect millions of truck and taxi drivers to learn new skills to become employable again?
Now, as a first observation on this question one should of course ask: what happened to all those agricultural workers when they became superfluous? What happened to the makers of horse carriages, buggy whips, carriage drivers, horse breeders, etc., etc., when the automobile came along? Consider for a moment the photograph below, which shows the Potsdamer Platz (square) in Berlin in 1930.
Berlin, Potsdamer Platz, 1930 – notice something?
(Photo © Bildarchiv Preußischer Kulturbesitz/ Herbert Hoffmann)
By 1930, the automobile had already thoroughly revolutionized transportation. And yet, in the picture above, we see three horse-drawn carriages in the middle of all this automobile traffic. This illustrates an important point the authors of the video are not considering.
Technological know-how is already far more advanced than is obvious in our daily lives. For instance, the driver-less car exists for quite a while already. Arguably, the technology required to build a working prototype of a driver-less car has been around for at least 25 years, if not longer. So why are truck drivers still safe? Why is the world not already full of driver-less cars?
What the authors of the video are not taking into account is that the limit to progress is actually not given by the state of our technological know-how. As we pointed out in a recent post on the use of mathematical knowledge in physics (see “Apriorism and the Natural Sciences” for details), physics as a theoretical science is in some respects decades ahead of the possibility to test its theorems by controlled experiments. It is very similar in many other areas of human endeavor: we know how to do certain things, but still cannot implement them on a grand scale yet. The reason for this is that the limit to progress is in fact the scarcity of capital.
Consider what would need to be done to reorient the entire transportation system toward driver-less cars. Obviously, such a system would involve building a sufficient number of such cars, as well as equipping roads with the necessary guidance systems. There are only two ways in which the vast amount of capital required to do all this can be obtained. Either enough new savings must be accumulated to allow for the creation of additional capital, or capital will have to be withdrawn from already existing employments. Note here that theentire chain of production activities involved in automotive production would require new investment.
When new technologies are deemed to have become economically viable and investment employing such new technologies is undertaken, often both additional capital from new savings will be employed and capital currently employed elsewhere will be put to different uses. How is it determined whether this can be done? The method all entrepreneurs must resort to is economic calculation: they must compare the cost of the required inputs with the expected proceeds from the output they will produce. It is however not enough to conclude that something is in principle viable (i.e., that the proceeds will exceed the cost of input). The new projects must also compete with all other current and planned employments of capital.
The reason why e.g. a new, better method of production is not immediately adopted by everybody is precisely the fact that it may not be sufficiently remunerative compared to already existing production processes. At the outset, entrepreneurs are faced with the structure of production as is, this is to say, the shape it has adopted as it has grown over time. Consider now an entrepreneur who wants to improve an existing method of production – i.e., he doesn't want to produce a new good, but wants to produce an already existing good more efficiently. Let us consider a simple product such as metallic bolts and nuts that can be used in numerous other production processes. A new machine has become available that would be enable an increase in output by 30% compared to the machines currently in use. The entrepreneur in our example erects a factory that is equipped with the new machine. Will all existing factories immediately throw out their old machines and replace them? They very likely won't. In fact, old production processes often remain competitive with new, more efficient ones for quite some time. Whether the new machine will be bought depends on the degree of its superiority.
Ludwig von Mises gives a simple example of this in Human Action:
“Let p be the price of the new machine, q the price that can be realized by selling the old machine as scrap iron, a the cost of producing one unit of product by the old machine, b the cost of producing one unit of product by the new machine without taking into account the costs required for its purchase. Let us further assume that the eminence of the new machine consists merely in a better utilization of raw material and labor employed and not in manufacturing a greater quantity of products and that thus the annual output z remains unchanged.
Then the replacement of the old machine by the new one is advantageous if the yield z (a – b) is large enough to make good for the expenditure of p – q. We may disregard the writing off of depreciation in assuming that the annual quotas are not greater for the new machine than for the old one. The same considerations hold true also for the transfer of an already existing plant from a place in which conditions of production are less favorable to a location offering more favorable conditions.
Technological backwardness and economic inferiority are two different things and must not be confused. It can happen that a production aggregate which from a merely technological point of view appears outclassed is in a position to compete successfully with aggregates better equipped or located at more favorable sites.
The degree of the superiority provided by the technologically more efficient equipment or by the more propitious location as compared with the surplus expenditure required for the transformation decides the issue. This relation depends on the convertibility of the capital goods concerned.”
(Emphasis added)
This is also why one can still see a few horse-drawn carriages in the streets of Berlin in 1930. It still made sense to employ them, in spite of the competition of the automobile. Not all of them were scrapped yet, as they were still able to render a valuable service. What the above really tells us is that already existing capital (note Mises' remark on the convertibility of capital goods, which is an important aspect of these deliberations) actually serves as a brake with respect to the speed with which technological innovations can be sensibly implemented.
Thus, even if we were to agree with the premise that one day, driver-less cars will replace current cars (it needs to be interposed here that one should never be too certain of what the future will actually bring), the fact remains that an army of driver-less cars cannot just drop from the sky. It requires scarce capital and also scarce labor to build them. It will take a long time to get from the current state of affairs to the future state of affairs when driver-less cars are ubiquitous. It is also quite possible that consumers will actually consider other employments of capital more urgent in the short to medium term and that as a result, this particular innovation will only be implemented on a large scale in the distant future. Even then, there will be a lengthy transition phase. Not all drivers will lose their jobs overnight, the process will be gradual. Hence there will be time for the labor force to reorient itself as well (this is precisely what happened in all other job categories that have disappeared over time).
The same applies to all the other examples presented in the video. We would also note, the fact that the number of jobs in the computer industry is smaller today than the number of jobs in transportation is actually proving anything. After all, the economy is not static. A decade hence, there may well be twice or three times as many jobs in the computer industry than today. There is no reason to assume that the number of jobs in this category today is in any way indicative of its future growth potential.
Labor as a Scarce Resource
Lastly, it should be mentioned that labor is obviously a scarce resource (otherwise it would not command a price). Is there anything that could alter this state of affairs? Technological progress need not be feared in this context. Today's high rates of unemployment in many developed nations are solely the outcome of interventionist policies, both in labor markets directly and indirectly via monetary policy and the booms and busts it produces. The excess of unemployment above the level of voluntary unemployment, we may refer to as “institutional unemployment” – it is caused by the fact that governments interfere with the market economy.
As long as there are unfulfilled human wants and as long as there is more land (in the widest sense) than labor, labor will remain a scarce resource (keep in mind that all capital originates from combining land and labor). In theory the possibility that the global population could grow so large that land becomes scarce relative to labor exists, but we are far from this point, and there are actually good reasons to assume that this point will never be reached. This is based on the purely empirical observation that population growth in highly developed nations tends to flatten out and even reverse.
As Rothbard wrote on this point in Man, Economy and State:
“[…] a crucial distinction between land and labor is that labor is relatively scarce. As a result, there will always be land factors remaining unused, or “unemployed.” As a further result, labor factors will always be fully employed on the free market to the extent that laborers are so willing. There is no problem of “unemployed land,” since land remains unused for a good reason. Indeed, if this were not so (and it is conceivable that some day it will not be), the situation would be most unpleasant.
If there is ever a time when land is scarcer than labor, then land will be fully employed, and some labor factors will either get a zero wage or else a wage below minimum subsistence level. This is the old classical bugbear of population pressing the food supply down to below-subsistence levels, and certainly this is theoretically possible in the future. This is the only case in which an “unemployment problem” might be said to apply in the free market.”
(italics in original, bolding added)
We cannot predict what the labor market will look like 20, 30 or 50 years hence. All we can say with certainty is that as long as the free market is allowed to operate – even in its current, severely hampered form – there will be things to do, leisure time and incomes will continue to increase in the long term, and the shape of the labor market will continue to change significantly.
Conclusion:
The future is always uncertain, but economic laws are time-invariantly valid. It is undoubtedly true that many of today's jobs will disappear over time. However, there will be jobs in the future we cannot even conceive of yet. Many of the job categories that have only come into existence over the past 30 years are highly likely to continue to see growth as well. Labor-saving devices have always caused people to fear for the prospects of employment, but these fears have always turned out to be misguided. Advances in robotics won't change that.
Charts by: USDA, Minnpost.com, NBER, First trust portfolios, suyt's space/FAOStat
