Home > Uncategorized > Badly confused economics: The debate on automation

Badly confused economics: The debate on automation

from Dean Baker

The media have been filled with accounts in recent years of how automation is displacing workers and threatening the country with mass unemployment. Even President Obama even made a point of warning about the dangers of mass displacement from automation in his farewell address.

This obsession is bizarre for two reasons. The first is a simple empirical point. In contrast to the concern about automation leading to massive displacement, in recent years the pace of automation has been extremely slow. Productivity growth, which is a measure of the rate at which workers are being displaced by technology, has averaged less than 1.0 percent annually in the United States over the last decade.

By contrast, it averaged almost 3.0 percent annually in the decade from 1995 to 2005. Productivity growth also averaged almost 3.0 percent annually in the long Golden Age from 1947 to 1973. This slowdown has not been restricted to the United States. Virtually every wealthy country has seen very slow productivity growth over the last decade. The United Kingdom even had several years of negative productivity growth. This is equivalent to workers were replacing robots: a situation where it takes more workers to produce the same amount of output.

So at a time when automation is proceeding at an extraordinarily slow pace we are seeing many policy types and politicians worrying about mass displacement from automation. That does not make a great deal of sense.  The other reason why the concern over automation seems misplaced is that it is directly at odds with how we talk about other areas of economy policy. To take an example that has recently been in the news, the Federal Reserve Board raised interest rates in the United States last month. It is widely expected to raise interest rates several more times in 2017.

The reason for raising interest rates is that the Fed is concerned that the economy is creating too many jobs. This will increase workers’ bargaining power, putting upward pressure on wages. A more rapid rate of wage increases will lead to more rapid inflation. To prevent this outcome, the Fed wants the economy to have fewer jobs.

But how can it make sense that, at a time when we are worried that automation is destroying a massive number of jobs, we also need the Federal Reserve Board to add to the job destruction by raising interest rates? If automation is leading to mass job destruction the Fed should not have to be worried about overly tight labor markets.

The same story applies to often repeated concerns about the demographics of an aging population. The standard story, which is repeatedly endlessly by the policy elite, is that we will have too few workers to support a growing population of retirees.

Apart from the basic demographics making no sense (we have always had a rising ratio of retirees to workers), the argument is 180 degrees at odds with the automation story. If automation is going to radically reduce our need for workers, then supporting a growing population of retirees will be no problem whatsoever. Incredibly, some of the automation scare story promoters simultaneously worry that we will have shortages of both jobs and workers.

In fact, the often voiced concerns about government deficits and debt also make no sense in the context of automation destroying jobs. What is the bad story if the government runs large budget deficits in a context where technology is hugely expanding our productive capacities?

The problem of budget deficits is supposed to be one where government spending is overburdening the economy, leading to either high interest rates, which crowd out investment, or runaway inflation. But if automation is hugely expanding our productive capacities then it should be able to accommodate the demand creating by large budget deficits. If automation is rapidly destroying jobs, the budget deficits should not be a problem.

There can be a story that automation shifts demand in the labor market, sharply reducing demand for less educated workers will increasing the demand for more skilled workers. This is theoretically possible, but there is no reason to believe that it describes the economy today or in the future.

While automation has displaced many jobs requiring less education it can also displace many of the most skilled workers. Developments in diagnostic technology are likely to make a skilled technician as good as the best doctor in assessing patients’ physical conditions.

Robots are likely to be able to perform surgery better than today’s best surgeons. And, with the Internet, there is no reason that any X-Ray or MRI scan should ever be read by a highly paid radiologist in the United States, as opposed to a much lower paid one in India. There is a similar story with lawyers and legal research and almost every other highly compensated occupation.

In short, there is no reason to believe that technology has any inherent bias towards destroying less skilled jobs while enhancing demand for the most highly skilled. It is likely to destroy jobs in both areas, although it is entirely possible that the most highly paid professionals will be able to use their power to block technologies that threaten their livelihood, as they have largely done thus far.

But of course this is not a problem of technology. This is a problem where privileged groups have the political power to be able to protect themselves from technology and the market. If President Obama and others warning of the threat of automation had made this sort of warning, they would be contributing hugely to the political debate.

As it is, their complaints about automation just work to sow confusion. It is very convenient for elites to blame the growth of income inequality on impersonal forces like the development of technology, but it is not true.

See article on original site

  1. February 8, 2017 at 3:46 pm

    “Productivity growth, which is a measure of the rate at which workers are being displaced by technology, has averaged less than 1.0 percent annually in the United States over the last decade.”

    You kind of lost me here. I see no reason to assume that productivity growth accurately measures jobs lost to automation. Wouldn’t goods have to be actually bought before they show up in productivity growth? When high-paying jobs are lost people buy less which more than counteracts any widgets per hour effect that would show up in corporate revenues divided by production-related wages.

  2. patrick newman
    February 8, 2017 at 3:46 pm

    However there will be heavy demand for the relatively small number of people who design, make, programme and maintain robots, assuming that robots will not beget robots for many decades – that’s the stuff of science fiction and the movies, or is it!

    • February 10, 2017 at 3:50 pm

      In 30 years Robots will be better at making robots than people are, assuming technology is allowed to continue developing.

  3. antireifier
    February 8, 2017 at 4:34 pm

    I usually find Baker’s material very helpful but not this time although it made me think.

    The first problem is his so-called empirical evidence about technology and automation. Predicting the future based on the past does not work very well. Empirically speaking it is often wrong. But allow me an indulgence on this. Near where I live, the local GM plant had 27000 workers on the lines and now has 6000 producing the same number of significantly more complex vehicles. This has happened in my lifetime. Which leads to another point. Namely tipping points. When technology happens it seems to have an exponential growth curve starting slowly and then exploding.

    There is a difference in technology that is software driven and automated technological changes in the past. The article also brought to mind chickens and money for me. I had a distant relative by marriage who automated the cleaning and packaging of chickens displacing many workers (saving their hands from cuts and other harms) but no computer software was involved. It was all cleverly designed hardware which he later applied to automated teller machines for dispensing cash. Now there is software involved in the cash dispensing but not when he was developing the mechanics.

    But more relevant to me is how these two events seem similar to what I can see on the (near?) horizon. UBER and driverless cars and trucks paired with just in time delivery. Tie these to the construction industry and you can easily see design software that is enacted in a modular factory where the components for homes are made by (3D technology perhaps), robots and then delivered to a factory by a driverless truck, unloaded and assembled by robots, placed on driverless trucks and delivered to a building site where they are assembled with a minimum of labour in a couple of days. Calculate the impact on labour of those.

    Then there is his citing of the Federal Reserve’s astounding ability to predict how things are going to go in the economy. Remember the 2008 Meltdown? Nice forecasting there. In any case the federal reserve is about protecting the interests of bankers not workers. In their mind workers are inflationary. They cannot see the other variables that are inflationary.

    So I get that the fed should be on top of this as Baker says, but there is an aspect related to the deficits and debts of governments raised by him where he misses again in my opinion. If what I have described above is correct, there will be a massive number of jobs lost quickly (exponentially?) which without better government planning (so much loved in the USA) will lead to a double whammy. Less taxes paid by unemployed workers needing significant aid packages.

    But all is not lost in the article. His comment about the privileged protecting themselves and Obama missing that factor is bang on.

    • February 10, 2017 at 3:53 pm

      Another data point on the exponential growth issue from mother jones:

      • February 12, 2017 at 5:17 am

        This way of comparing the human brain and computers in terms of computing power does not consider the quality and virtue of calculations. The killer robot of the movies may be capable of 2.88 x 10^17 calculations per second. But if this only serves to make the robot a faster and more thorough killer, the computer brain of the robot is still much inferior to the average human brain.

      • February 12, 2017 at 5:59 am

        A lot of AI skeptics have become believers since computers beat Lee Sedol in Go. Many thought that computers would take until 2050 or longer to reach that milestone based on Deep Blue-style calculation densities. The milestone was reached by changing the way computers think to be much more like the way we do it.

        I understand a little bit about how these changes were implemented and I am a believer.

      • February 12, 2017 at 8:09 am

        I’m not a skeptic about AI. I don’t believe it’s possible until humans have a much fuller and clearer understanding of human intelligence. Humans invented intelligence about 20,000 years ago. It was in the fullest sense a great fantasy about how Homo sapiens live, address the world, and finds its way. Since we invented it, we need to answer the questions about what it is we invented and what we want to do with it. Once we can do that then AI may be possible. But even then AI must remain a pale and incomplete copy of what human intelligence was invented to be (once we believe we know what that is, if ever).

  4. February 8, 2017 at 5:10 pm

    How about taxing the use of robots in order to support unemployed workers?
    The next step is to implement a guaranteed income for all based on robot sweat.
    Then we will all be aristocrats!

    • February 10, 2017 at 3:55 pm

      “How about taxing the use of robots in order to support unemployed workers?”

      Keep it simple. 20% tariffs to prevent capital flight followed by a 20% revenue tax on any company with a greater than 20 ratio between highest paid and lowest paid including subsidiaries and captive suppliers. The 20/20/20 plan for 2020.

    • Rhonda Kovac
      February 14, 2017 at 10:44 pm

      Excellent point. I’m sorry. Mr. Baker, but automation is progressive and unstoppable. The problem with it is that profits from machine labor — together with profits from almost all other economic productivity gains — go overwhelmingly to the rich. Some sort of redistribution — such as would be involved in a guaranteed income — is an essential part of the solution.

      I think we should be thinking even deeper, to challenge the longstanding notion of anchoring income to labor — the idea that everyone has to make his or her own living. With advances in automation this is becoming more and more out of reach.to a greater segment of the population., Work will of course continue to be a part of our economy, but the fundamental economic principle should be more along the lines of “from each according to ability, to each according to need.”

      • February 15, 2017 at 7:15 am

        Ronda, completely agree. The problem is not labor and work. The replacement of humans by robots in some areas of work is really not even an economic concern. It is a moral and ethical concern. Does the displacement of humans violate moral and ethical standards that give humans the right to expect a useful life that makes positive contributions to the community? Or can this be taken away by robots?

        The relationship between work and welfare is seen in all the history humanity, as described above. That one’s physical place in society and personal/family welfare is dependent on that work in any monetary sense is not. That many today believe it is simply is an historical accident. It’s an expansion of a few notions of the Enlightenment and Industrial Revolution. And even here it’s contradicted by statements from the same sources that the Enlightenment and Industrialization would bring bounty and greater leisure for all humans, regardless of work or abilities. Marx clearly recognized this. Thanks for the quote from Marx. Such ideas are not popular in much of the world today. Certainly not in the US. Curious, since Marx was paraphrasing the words of Christ.

      • February 15, 2017 at 7:16 am

        Rhonda, please excuse my misspelling of your name.

      • Rhonda Kovac
        February 15, 2017 at 12:47 pm

        Ken, Thanks for complimentary remarks. FYI, “from each according to ability, to each according to need” was a popular European socialist saying antedating Marx, which he requoted.

    • February 16, 2017 at 6:20 am

      Am aware of the history of the saying. Just like to get Marx’s name out there when I can. One way to get a better read on people is to watch the reactions.

  5. Craig
    February 8, 2017 at 6:58 pm

    Sorry, amongst other economic orthodox thinking the article is cluelessly hidebound to the current monetary paradigms of debt, loan and for production only.

    • February 8, 2017 at 10:01 pm

      Indeed, the speculative economy does not even enter into the discussion.

  6. February 9, 2017 at 5:59 am

    What is the relationship between money, on one side, and on the other side production, employment, wages, and economic inequality? This is how economists see these relationships, I think. Money is needed to fund production (salaries, factories, equipment, etc.). That money comes from investors. These investors expect their money to be returned to them, along with an additional amount of interest on that money. Workers (either the human sort or the non-biological sort) are needed to produce. The workers are paid, if biological, and maintained if non-biological. The pay to the biologics is as small as possible, set based on what others will expect in pay to preform the same tasks. Maintenance and depreciation of non-biologics in measured in terms of pay for biological workers. Those hiring the workers (biologic) want high productivity per worker. As that productivity per worker increases, workers receive only the minimum amount of that extra income to keep them working. Non-biologics work based on the schedules set, stopping only for mechanical repairs or replacement. Since this arrangement leaves only the owners of the productive facilities in charge of distribution of ALL the money from the production process, this inevitably leads to the division of richer people (owners and investors) and poorer people (workers). This last result changes only if the owners and/or investors choose for it to end. Politicians generally support this process and its results. As do most economists.

    Did I get it right?

    • February 10, 2017 at 4:25 pm

      “Did I get it right?”

      I would agree with much of that, but I would add a few details.

      1) The fed, when setting interest rates, does not nullify the law of supply and demand and replace it with the law of the fed, but rather it works within existing supply and demand. Supply of dollars for investments comes from the supply of dollars in the hands of rich people. Supply of dollars for return on investment comes from the supply of dollars in the hands of the unsatiated (bottom 95%?). Interest rates and stock and real estate ROIs are all set by the inequality policy of the country and are determined today primarily by the fact we live under the Reagan tax cuts, which are driving all types of ROI steadily toward zero independent of what the fed may want.
      2) As the bottom 95% have less and less money to spend from wages as a percentage of GDP, how is economic growth still possible? The answer lies in supply-side leakage in the form of housing and stocks to the asset holders among the middle class. Obviously this avenue loses leverage as the ratio of assets held by the middle class to assets held by the satiated class decrease.
      3) Inequality feeds into an asset paradigm of lower and lower percentage volatility in the asset markets until an earthquake comes. If the earthquake is manageable we get the Great Depression. If not, we get the French Revolution.

      • February 12, 2017 at 4:49 am

        Thanks, jeff1089. I take your comments as arising from an economics perspective. I take them that way in any event. In those terms, they’re helpful for me. Thanks.

        Now an anthropological question. What survival value, if any do the ways of life you describe have? In other words, from an evolutionary perspective what survival advantages do these ways of life provide?

    • February 12, 2017 at 6:18 am

      “What survival value, if any do the ways of life you describe have? In other words, from an evolutionary perspective what survival advantages do these ways of life provide?”

      I am not sure I understand the question, but the biggest untold story in behavioral selection is plasticity. Our boldness is determined largely by our treatment early in life. As a metaphor, when girls are young they establish norms of what is attractive in men that are more powerful than most later inputs. Similarly, when humans are young they establish internal norms on how much they trust their own opinions over the opinions of others and how much empathy they feel. For example, ability to follow the scientific method is not determined by native ability so much as ingrained hubris that is shaped by position in society and treatment during the first few years of life. Similarly, empathy is magnified by the experience of frustration at an early age. People who don’t experience roadblocks in the way of getting what they want from a young age miss critical development steps in the development of empathy. It is harder, but still possible, to acquire this perspective later in life, just as it is still possible to learn a language later in life, but the odds go way down from a circumstance/motivation perspective.

      A genotype exhibiting this plasticity yields individuals that had a lot of offspring when they were lucky enough to be chiefs while not going extinct when put in the serf role. Very likely it was a pre-requisite for the development of any sort of “social” animal.

      Unfortunately today this means that when someone is groomed to be CEO, they will likely be empathy-deficient and will remain so until something in life gives them hard lessons.

      We have amplified this with changes to rules enabling LBOs with less than 10% down. Empathy in most cases within mature industries of the business world represents value waiting to be unlocked. This drives propaganda in the overall population with effectiveness dependent mainly on how far we are from the last major set of crises. From this perspective the Kondratiev cycle can be thought of as being like an LC-circuit in electronics. The time between “zaps” is a function of the constant rate of charge built-up and the nearly constant dielectric breakdown point (substitute people remembering the last crisis dying off for charge buildup and Minsky crisis for capacitors becoming conductive to complete the analogy).

      • February 12, 2017 at 7:59 am

        This fairly standard psychology, as that is understood over the last 100 years or so. It’s not the only possible of psychology that could be. In fact, psychology did not come into being as a necessity. It was an accident of social evolution. Your comments fit with one particular social evolutionary path. There are countless others down which social evolution for Homo sapiens might have moved.

        I’m not referencing psychology (any version) but rather social evolution. My questions are about social and in some instances biological evolution. As with other species Homo sapiens exhibits random variations in its biological structure and functions. Those variations that enhance survival (biological) likely continue. Those that do not generally die out. Social evolution is not quite so clear. Sometimes variations in social structure or functions continue even when they harm survival chances. So, do the cultural changes you list (a belief in supply/demand, creation of unsatiated groups, earning a return on investment, inventing and owning assets) provide any survival value for those who accept and live under them?

    • antireifier
      February 15, 2017 at 3:37 am

      Do cows, pigs, chickens, salads, etc. as biological entities get paid? News to me. And not all humans get paid. Slavery still exists.

      • February 15, 2017 at 6:56 am

        As the western economies are currently organized, if they work and are not slaves they are paid. Not always in money or money substitutes. Sometimes they are simply provided food and shelter sufficient to stay alive. Most slaves are not paid. Although, if you check I think you’ll find that in ancient Greece and Persia, among others slaves were often paid, and could buy their way out of slavery.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s