Skip to content

The Truth About Automation, Jobs And Prosperity

2021 July 11
by Greg Satell

Since the early days of the Industrial Age, there have been competing visions about the effects of technology. Mary Shelley’s Frankenstein monster has been the model for technology gone awry while, ironically, Karl Marx’s vision of techno-utopianism has been adopted by avid capitalists and Silicon Valley libertarians.

At the heart of the debate lies automation. When we use machines to perform tasks formerly done by humans we unlock multiple effects. The most obvious, of course, is that somebody is out of a job, less obvious are how technology affects productivity and creates new industries, which we hope will create more and better jobs.

Yet there’s no guarantee that technology will raise all boats. Markets are complex ecosystems and things can’t always be broken down into simple, linear relationships. What we can do, however, is get a better understanding of how automation affects our society. Only then can we build a consensus about what we want outcomes to look like and work toward them.

How Automation Affects The Economy

In 1900, 30 million people in the United States worked on farms, but by 1990 that number had fallen to under 3 million even as the population more than tripled. So, in a manner of speaking, 90% of American agriculture workers lost their jobs, mostly due to automation. Still, the twentieth century is seen as an era of unprecedented prosperity.

The idea that putting farmers out of work can potentially be a good thing is nothing if not counterintuitive. To get a handle on how that can happen, we need to look at three separate forces automation unlocks: A displacement effect, a productivity effect and a reinstatement effect, each of which can affect the economy in myriad ways.

To understand how these effects help shape an innovation economy, let’s look at those farmers again. During the late 19th and early 20th century agriculture was transformed from being largely driven by a combination of animal power and back-breaking human labor to a heavily mechanized industry. For example, it’s been estimated that in 1940 alone, tractors saved 1.7 billion hours of human labor.

Obviously, technology displaced a lot of people who had been working on farms. However, the increased productivity made farmers richer and allowed them to pay workers more. That allowed people who grew up on farms, such as a young Henry Ford, to begin tinkering with machines and create entire industries and new types of labor.

Since the beginning of the industrial era, automation’s productivity and reinstatement effects of have outweighed displacement effects. Economists, almost as an article of faith, tend to assume that the same will be true of new technologies. However, there is no evidence that it is necessarily true and, in fact, we may be seeing signs that it is not what’s happening now.

The Rise Of The So-So Economy

It’s been clear for some time now that we’ve been in the midst of a second productivity paradox. The first one, which lasted from the early 1970s to the mid 1990s, saw diminished productivity gains amid increased investment in information technology and prompted economist Robert Solow to note, “You can see the computer age everywhere but in the productivity statistics.”

In 1996, with the rise of the Internet, productivity growth began to boom again but then disappeared just as abruptly in 2004 and hasn’t returned since. Despite the hype surrounding things such as Web 2.0, the mobile Internet and, most recently, artificial intelligence, productivity continues to slump.

Part of the answer may have to do with what economists Daron Acemoglu and Pascual Restrepo refer to as “so-so technologies,” such as automated customer service, which produce meager productivity gains but displace workers nonetheless. Such technologies could also alter the relationship between displacement, productivity and reinstatement effects.

Consider an airport bar where ordering has been automated through the use of touchscreens. It’s hard to see how, given the high rent, food preparation and other costs, this technology would have a dramatic effect on productivity akin to, say, replacing a horse with a tractor in an agricultural economy. Still, the displacement effect is substantial.

In fact, Acemoglu and Restrepo argue that a large-scale version of this phenomenon has been occurring since the late 80s. Digital technologies, to a large extent, have created even larger displacement effects than earlier innovations, but have not had the same offsetting productivity and reinstatement effects as earlier technologies.

The Problem With Trying To Train Our Way Out Labor Displacement

When a task is automated it is also commoditized, which creates opportunities to create value at a higher level. For example, automated bank teller machines have led to an increase in bank tellers. Instead of helping people do routine tasks like deposits and withdrawals, these tellers are doing non-routine tasks, like dispensing financial advice and resolving problems.

So it seems logical that investing in worker training as a society would be a viable solution. Instead of firing all the tellers and replacing them with teller machines, the banks trained them to give expert advice and provide seamless customer service. In that scenario, everybody wins. Yet this particular narrative can be very misleading.

In fact, Peter Capelli at the Wharton School argues that the entire notion of a skills gap in America is largely a myth. One reason that there is such a mismatch between the rhetoric about skills and the data is that the most effective training often comes on the job from an employer. It is augmenting skills, not replacing them, that most often creates value.

In fact, when you take into account “so-so innovations” such as automated customer service, the rationale begins to melt away. Remember, these technologies provide minimal productivity gains, which leaves little capital to finance reinstatement effects. In other words, there’s little evidence that putting check-out clerks out of work will create enough value to finance entire industries for which those clerks can find work as graphic designers.

We should also note that, in the United States at least, competition has been declining for the last two decades. So in many industries, such as cable TV, healthcare and airlines, firms have an incentive to automate even if it diminishes customer satisfaction while, at the same time, there is increasing evidence that automation is diminishing wages for most workers.

Establishing A Human-Centered Economy

What’s crucial to understand about automation is that technology does not exist independently from us. As Martin Heidegger suggested in the middle of the last century, it comes into being through a process of revealing and building. First, scientists reveal certain truths about the nature of the universe and then we channel those forces to build things we see value in.

When we automate things like customer service, we are making a series of value judgments. Replacing a server at an airport bar, for instance, says something about how we value interactions with a friendly server. It also frequently says something about how we value competitive markets, since such establishments usually have a monopoly over concessions.

What’s becoming clear is that we can’t simply leave our fates to the impersonal whims of market and technological forces. The evidence shows that innovations that displace workers do not necessarily make us better off. In fact, we have strong reasons to suspect that many of these technologies impoverish our society and corrode our culture.

Put simply, we need to make better choices and rebuild a human-centered economy based on dignity, in which we treat people as ends in themselves, rather than as means to an end. Technology and markets were, after all, created by humans to serve people. That is their purpose and should be, by any reasonable analysis, the measure of their value.

– Greg


Image: Pixabay


7 Responses leave one →
  1. July 11, 2021

    This is really really good. So much “so so” innovation with little increased productivity and far greater increases in aggravation. You can always sell owners/managers on any technology that will get rid of pesky expensive employees. Just never mention the downsides. Agile isn’t a software development system we have adopted, it is an ideology at my current company. Also, automation allows centralization of ownership… quite a Marxist point. That very likely slows monetary “velocity”, hurting the entire economic system.
    You know that as a biologist, cum developer, cum philosopher, cum derelict, I take a different view. I ask what does a person require to survive….with the marketing removed? When are we able to produce all that we need with the technology we have?
    Also as a biologist, I would posit that what you are observing is an outcome of Darwinian win-lose drives and strategies for competition and dominance. Those would be in direct conflict with human strategies of a … “human-centered economy” or pretty much anything else that is a win-win human created strategy.

  2. Gerardo permalink
    July 11, 2021

    You wrote: ….. Robert Solow to note, “You can see the computer age everywhere but in the productivity statistics.”
    What Solow did not see in the productivity statistics is the abuse of taxation and the impact of wasteful government spending and regulations by the government.
    These are the main factors in the fall in productivity. Only a Nobel laureate can be stupid enough not to understand the enormous and tangible increase in productivity due to technological innovation since the seventies (floppy disk, cell phone,VCR,digital camera, microcomputer. …).

  3. July 12, 2021

    Thanks for your input Michael. I always look forward to hearing your thoughts.

    – Greg

  4. July 15, 2021

    Dear Greg, thank you for explaining the so-so technologies here. I also found the three effects (displacement, productivity, reinstatement) to be interesting. Do you think this is the same as the job destroying, job enhancing and job displacing categories used by WEF and others,

    Best wishes,


  5. July 15, 2021

    Thanks Shawn. I think that’s a little different. If I’m not mistaken, those reflect different categories of technologies. These are effects of technology that economists can measure (see paper linked into the article).

    So, for example, a single technology like a tractor could have all three effects. Tractors lessened the need for farm workers (displacement), allowed Henry Ford to make enough money to be able to inker around (productivity) and start a new industry which created new jobs (reinstatement).

    I hope that’s helpful.

    – Greg

  6. August 15, 2021

    Great post. I would quibble that the language used to describe farm labor (eg. “back breaking”) is a value judgment made from a time in which such work is unjustly devalued, and it’s benefits grossly overlooked. Sure, we’ve increased agricultural output in terms of sheer tonnage, but at what human and ecological cost? We’ve poisoned our soil and waterways, and we’re exceedingly more anxious and unhappy. As an artist and writer, I understand the benefits… but I am suspicious of the idea that it had to be done the way it was – with toxic chemicals and destructive machines. Until we face the way our past errors have led to our present misconceptions, we’ll never be able to make the creative leap required to envision a better way forward.

  7. August 16, 2021

    Good points. Thanks Josh.

Leave a Reply

Note: You can use basic XHTML in your comments. Your email address will never be published.

Subscribe to this comment feed via RSS