mead cohen berger shevtsova garfinkle michta grygiel blankenhorn
$15 Minimum
The Robot Subsidy Act

A new report from McKinsey confirms what we have known for a while: The types of jobs most likely to be affected by the minimum wage hike are also those most vulnerable to automation. Business Insider reports:

Do you spend a lot of time on predictable physical work, data processing, or data collecting? We’ve got some bad news.

A McKinsey report from Michael Chui, James Manyika, and Mehdi Miremadi says analyzing work activities rather than occupations is a more accurate predictor of automation.

Unpredictable physical work, stakeholder interactions, and applying expertise are less susceptible to automation. Managing others is least susceptible — so no robot bosses yet. […]

Spoiler alert: it’s not looking good for accommodations and food services and manufacturing.

As $15 per hour wage floors phase in in cities and states across the country, companies will have even stronger incentives to invest in robots. The job losses for less-skilled workers will accelerate. And while this policy will certainly harm struggling small businesses, it is very much in the interest of certain large corporations. Silicon Valley technology executives and venture capitalists will reap the biggest benefits from what is essentially a subsidy to the information technology sector.

To be sure, the trend toward automation would happen with or without radical minimum wage hikes—and, in the long run, the elimination of mindless, routine tasks is a good thing. But at a time of economic transition, when labor force participation is dwindling, we have no reason to accelerate what is in the short run a painful and disruptive trend. Policymakers genuinely interested in reducing inequality and raising standards of living for those at the bottom should look to other areas—like tax credits, housing, immigration, and job training—to raise workers’ incomes without replacing them.

Features Icon
show comments
  • LarryD

    The author assumes that the minimum wage hike advocates actually give a damm about the little people.

    “By their fruits you shall know them.”

    • Blackbeard

      The question to ask is not, “Will this policy help poor people?” The right question is, “Will this policy help Democrats win elections?”

      After we enact a national $15 an hour minimum wage, and we will enact it, we will see an increase in poverty and joblessness among the weakest in our society and that will particularly include minorities. The Democrats will then blames this on the heartless Republicans and their stubborn unwillingness to run even larger deficits, as though $600 billion a year isn’t enough. The media and academia will, of course, back the Democrats.

  • f1b0nacc1

    The $15/hr minimum wage will be great for people like me (part of my work involves robotics), but it will be disastrous for those that actually make minimum wage.
    Automation (more than simply robots) can already do a great deal of the work done by minimum wage workers, and as the technology gets more capable (and it is doing so very, very quickly), the fraction of the work that can be done by automated systems will grow. No Ned Lud is going to stop this, and no set of policies can change this. No- and low-skilled work occupies a rapidly shrinking area of opportunity.
    With that said, however, automation is capital-intensive (read: expensive), and often requires process reengineering to be practical. This means that there are significant opportunity costs involved with implementing automated systems, and such costs have acted to prevent it from being adopted even in areas (notably fast food) where it could be widely used. A higher minimum wage will have a very positive (for those like myself who stand to benefit from automation) by significantly reducing the opportunity costs associated with adopting automation by increasing the costs of not doing so. This is simple mathematics, and whether one likes the results or not, it is not open to debate. Thus a higher minimum wage (particularly one as ridiculously high as being proposed by the $15 twits) is simply making it easier for enterprises to take the plunge and introduce technologies that they would have otherwise been likely to otherwise delay.
    The impact of this cannot be overstated. Under a normal free market, the process would take place relatively slowly at first, rapidly accelerating once a tipping point was reached (i.e. when sufficient number of enterprises had automated, and a critical mass of knowledge and experience dealing with automation was acquired), but with this serious market distortion, it will occur much faster, often before a widespread base of accumulated experience will be available. Worse, these changes will happen far faster than the glacial organs of policy will be able to catch up (if indeed it is even possible to do so, I have my doubts), and certainly before societal norms can.
    Translation: it is happening, it will now happen far faster, and it will not be pretty for anyone other than those who make their living introducing it….

  • Pait

    There’s something fishy about the line of reasoning. Serious automation is expensive. It tends to be used in industries that employ workers with reasonably high wages, and rarely to displace minimum wage workers. The idea that a minimum wage increase will lead to significant investment in automation is questionable at least.

    • Spencer

      “Serious automation… tends to be used in industries that employ workers with reasonably high wages”

      Then you admit that there is a direct correlation between wages and automation. So if wages rise, so will automation. This is not difficult.

      I think you could have responded to yourself with an ordinary pocket calculator. A $15-an-hour full-time worker costs over $30K a year in wages. Federal regulations alone add about another $10K, then state and local regs even more (sometimes much more). How expensive do you think a hamburger-flipping machine would be? They’re not gold-plated, you know. And while they may require repair and occasional replacement, they don’t come in late, call in sick, steal company supplies, and their work is absolutely consistent. If they need to be replaced, it can be with another model of exactly the same (or higher) quality, no HR department, background checks, or want ads need.

      An employee that may be worth keeping on when you’re spending ~$30K a year in wages and regulations may very well be disposable at ~$45K. With automation, you pay for the machine once and then only pay for maintenance and electricity. Automate the position, and in five years you’ve saved over $200K in labor costs.

      • Pait

        In the long run, the correlation is the opposite: as automation increases, wages rise, because of productivity.

        My point is that the easy ways to automate low paying jobs have been implemented. Many low paying jobs are quite difficult to automate further. Not so much with high-paying jobs, where the benefits of automation justify larger investments.

        Your numbers seem to have been plucked out of thin air, and are not really a starting point for a longer discussion. It is not the case that a “hamburger-flipping machine” would be enough to replace too many jobs. You need to think a little harder about the tasks involved in the fast food business and the types of robots or whatever that are within the possibilities of current engineering to understand the trade-offs involved.

        • Spencer

          “In the long run, the correlation is the opposite:”

          No. Higher labor costs lead to greater automation. The correlation is exactly as I said. If you increase the price of something, people will seek a substitute for it. This remains true whatever time-frame you choose. You may seek other correlations, this one remains, and it’s the whole point of this article.

          “as automation increases, wages rise, because of productivity.”

          Well, I agree that this may be so in the long run, at least theoretically. But, as someone said, people don’t eat in the long run, they eat everyday. And doubling the minimum wage will put a lot a people out of work for years, until price levels adjust.

          “Many low paying jobs are quite difficult to automate further.”

          I agree that there are low-skill jobs which will not be automated soon due to technological constraints or consumer sentiment. I will grant that there may even be certain particular jobs which can never be automated (although “never” is a long time, so I’m skeptical).

          But I presume many low-paying jobs are at the point where automation is feasible, but labor costs don’t yet justify the cost of automation. Doubling the minimum wage will put them over the edge.

          You say only highly paid occupations justify the cost of developing automation to replace. This is flatly incorrect. Occupations which are likely to be automated are those whose workers account for proportionally high labor costs collectively, whatever their wage levels. A job which includes a large number of workers who are paid low wages is a bigger target for automation than one which employs a much smaller number of higher paid workers. Then double the amount you pay the former…

          “You need to think a little harder about the tasks involved in the fast food business”

          I have worked in the fast food industry, albeit a long time ago. As you advised, I’m thinking really hard, but I’m having trouble thinking of physical tasks I performed which could not be decomposed into discrete motions which a simple mechanism or series of mechanisms could replicate. If you have those mechanisms, all you need is the software.

          “and the types of robots or whatever that are within the possibilities of current engineering to understand the trade-offs involved.”

          I once heard a wise saying: Don’t say something can’t happen until you’ve checked to see whether it has already happened.

          As for the numbers, a starting point for discussion is exactly what they are. The regulation costs are certainly no more than a ballpark figure. As for the wages, once again, calculators are your friends.

          • Pait

            The causation between productivity from automation and salaries is exactly as I stated.

            The reason why it is difficult to further automate low paying jobs is that low-hanging fruit have been picked. Developing automation tools is expensive. I know you could come up with a “machine to flip hamburgers”, or a “robotic vacuum cleaner”, but the sequence of operations needed to serve fast food or clean a house are difficult to replace. Many low paying jobs are in services that are actually quite complex, such as care for disabled or older people.

            As for the level at which higher minimum wages lead to loss of jobs, it is hotly debated. Surely there exists some level at which some jobs disappear, but what it is, we don’t know precisely. Moreover, there are macroeconomic phenomena that increase demand, thus employment, when wages rise. You write with a certainty that does not exist.

    • Brett Champion

      That might be true historically, but that doesn’t mean it will be true in the future. As robots become cheaper and better, the latter of which is happening at an accelerating pace, the cost-benefit analysis will shift inexorably in favor of automation and against human workers.

      The issue raised by the post, however, is still up for debate: Will the raising of the minimum wage to $15/hr. cause a significant enough of a shift to automation in the short-run to cause more harm than good for the targeted class of workers? My inclination is to say that it won’t. Though once things like this get going they sometimes ramp up much more quickly than what you would expect.

      I still think, however, that a national minimum wage is a bad idea, not to mention that the imposition of national labor regulations is unconstitutional, but that’s not on point to the issue here. As a tool for poverty alleviation, the minimum wage is a blunt instrument that causes considerable collateral damage to both the targeted class of people and the economy in general. The minimum wage, if set above the natural minimum wage (and what would be the point otherwise), has the effect of making some people unemployable because they are too expensive for an employer to take a risk on or because the minimum wage is simply set above an amount of economic value the person can produce. The people who get locked out of the labor market in this way tend to be the ones most likely to suffer severe, long-term consequences from not being able to land a job: the young, the mentally disabled, former convicts, etc.

      A national minimum wage also distorts regional economic patterns. That is to say that it effectively locks poor areas of the country into staying that way. After the Civil War, the South was an economic basket case, and it stayed that way for quite awhile. What allowed parts of the South to eventually escape that condition was the ability to attract factories from the rest of the country in part because of the lower labor costs that existed there. Areas of the South that had relatively productive workers and good locations relative to key transportation nodes were able to attract investment, which was how they started on the road to economic recovery. That’s not to say that areas like Appalachia can necessarily make the same leap, but a $15/hr. minimum wage effectively makes them no-go areas for any business that doesn’t have to be located in the area to perform its services or make its goods. It also means that any person with any ability and ambition must leave the area if he is to have a job commensurate with his skills.

      And as the post points out, if poverty alleviation is the real goal, then there are much better options available than a minimum wage. These options, especially tax credits, can be used with less distortion to the labor market and national economic activity.

      • Pait

        I agree that minimum wages should take into account local conditions. That is probably why the movement is being led by states and cities where demand for labor is stronger, such as California, the Northwest, Minnesota, and the Northeast.

        You speak with a lot of certainty about the future, while not being so sure of the past. I’d say this is unwarranted. Current automation technologies are making faster progress with middle and higher income office tasks. Automation of the lowest paid jobs is becoming more difficult, because of technological trends, leaving aside the fact that the returns are more limited because of the lower wages themselves.

© The American Interest LLC 2005-2016 About Us Masthead Submissions Advertise Customer Service