Automation and Employment: A Complicated Relationship

“Machines today can prepare your food, clean your house, and build a whole lot more machines. They’re out of control, coming to take our jobs and impoverish people—or are they?”

By David Behrens

     A bad habit of many of today’s commentators and even economists is to pay stringent attention to one economic variable and disregard how others behave in the long run. There is no better example of such a variable than employment; the idea of every member of society having a job is so appealing to people that even some economists prescribe solutions which prioritize this facet of national economies too heavily. It’s a useful political crutch; talking heads have gotten away for centuries with false rhetoric about government creating jobs, or more recently, the notion that jobs can be ‘stolen’ from people.

The Source of Confusion

     Likewise, the concept of automation has taken center stage in rhetorical debates about employment. Of course, we take it as a given that machines are used to consistently and efficiently perform tasks, many of which people could do themselves. It is stylishly philanthropic for the modern layman to cry out against the implications of permanently replacing human labor with machines. In fact, it must alarm all of us to imagine that one sudden mechanical innovation could permanently render obsolete the skills of thousands of good working people, stealing away their livelihood.

      However, taking the opposite view to its logical extreme is yet more frightening. The great French economist Frederic Bastiat once poignantly illustrated this point with his satirical essay “A Petition,” (1845) in which his narrator advocates for the government to extinguish the sun, because it takes away business from candlemakers. Now, no machine is the sun, but members of the global workforce aren’t mere candlemakers, either. Economists often fall prey to fortune-telling and dogma, but invisible mechanisms of human action have unpredictable, long-reaching effects on our societies. We can at best, consider the direct effects of automation and remain conscious of those which are not so obvious.

What is Employment?

     Employment, like many other concepts, is somewhat meaningless on its own (Bastiat 1850). This seems a daunting assertion to many people, as the concepts of hard work and career life have long dominated human culture. But it’s true; if resources were not scarce, and no effort was required to distribute them, labor would be entirely unnecessary. Humans work so that they can produce goods and services for everyone to enjoy. They are compensated based on the impact that they as a unique individual can generate for everyone else.

The Tradition of Human Progress

     At this point in time, there is one thing that humans produce that can never be produced by machines alone: innovation. To this end, automation has proven very beneficial in an interesting way. Innovation is considered the primary driver of growth in highly developed nations like the United States, whereas developing nations grow because of increases in capital stock and productivity (Friedman 1980). Automation in developed nations saves time and effort that humans in turn can use to invent new things to make life more easy and enjoyable. Eventually, new technologies make their way into developing nations, helping growth accelerate with increases in productivity.

Crucial Effects in the Short Run

     Classical economics demonstrates that in the long run, industries adjust, and workers will shift to jobs in other fields (Say 1803). However, large shocks to the labor market may have more severe consequences; the more workers lose their jobs simultaneously, the larger the ripple effect could be. The negative effects include decreases in consumption from those workers, as well as a possible surge in payouts from unemployment insurance. Positive effects follow as well; companies that can automate will cut costs, possibly allowing them to decrease prices and encourage consumption of their products. As I mentioned, consumers’ newfound disposable income may encourage some investment in innovative technologies and lead to economic growth.

     The main problem with all of this is that the negative effects come first, are more noticeable and can be politicized. The long-run benefits, however, manifest themselves more subtly over a long period. The initial panic can grow like a virus; if consumption suddenly decreases, production will as well, so there will be a little bit more unemployment, and so on. That is why the severity of short-run changes is important; this is a textbook example of Keynes’s theory of the trade cycle.

     Luckily, there are some safeguards against such a drastic level of structural unemployment and its implications in the short run. For one, it’s generally easier to automate production in sectors that use primarily unskilled labor (Robinson 2014). In these cases, unemployed workers aren’t liable to have wasted years of vocational training or specialized experience. This makes it easier for them to assimilate into different sectors. Additionally, the automation process is often gradual, and many large companies have positions for employees to move to instead of being laid off entirely.

     Still, the invisible effects of unemployment and those of automation cannot be taken for granted. They aren’t quantifiable, and have far-reaching effects of their own. If we are to see even temporary unemployment as too high a cost to offset the benefits of automation, it becomes another matter to properly regulate the advancement and use of innovative technologies. In the end, there is nobody who can predict all the consequences of any economic phenomenon. Our responsibility as economists and ordinary people alike is to look for hidden information whenever possible, recognizing that we cannot hope to completely understand it.



Bastiat, C. F. (1845). The Candlemakers’ Petition. In C. F. Bastiat, Economic Sophisms.

Bastiat, C. F. (1850). Ce Qu’on Voit et Ce ne Qu’on Voit Pas. Paris, France.

Friedman, M. (1980). Free to Choose. Chicago: Harcourt.

Robinson, A. (2014, October 22). Manufacturing Technology. Retrieved from Cerasis:

Say, J.-B. (1803). A Treatise on Political Economy (6th ed.). (C. C. Biddle, Ed., & C. R. Prinsep, Trans.) Philadelphia: Lippincott, Grambo and Co. Retrieved 2017