Technology and “Good Jobs”

NOTE: The Growth Economics Blog has moved sites. Click here to find this post at the new site.

I got a number of comments and e-mails regarding a recent post on technological change, jobs that produce goods, and “good jobs”. This is a follow up meant to clarify some points and solidify others.

  1. The entire point of my post was to say that the exact tasks people do is unrelated to whether they have a “good job”. Working in manufacturing does not make a job “good”, and working in services does not make a job “bad”. Yes, “good” and “bad” are fuzzy terms.
  2. I’m not denying that technology is replacing manufacturing jobs. It is. It will. We may well end up with robots making everything. If so, then you want to make sure that the jobs that people do have are “good jobs”.
  3. Service jobs are not, even to a first approximation, poor people doing things for rich people. So no, we won’t run out of jobs because rich people can only get so many massages or restaurant meals. The vast majority of workers in the US for the last 60 years have been non-rich people doing service-like things for other non-rich people. [Teachers, cops, firemen, nurses, waiters, store clerks, everyone in HR, everyone in accounts payable, secretaries, receptionists, every computer programmer, truck drivers, warehouse workers, chefs, everyone who works on any TV show, record, or movie, claims adjusters, insurance agents, financial analysts, everyone at your local bank, your IT guys, everyone working in state or federal government, priests, librarians, florists, pizza delivery guys, photographers, personal trainers, dietitians, optometrists, dentists, physical therapists, veterinarians, security guards, dishwashers, hostesses, exterminators, HVAC workers, plumbers, electricians, roofers, rodeo clowns, pit bosses, morticians, barbers, day-care attendants, real estate brokers, airline pilots, car mechanics, flight attendants, taxi drivers, and yes, even used car salesmen. Just to give a few examples.] We are very good at finding things to do for each other. We’ll continue to be good at that
  4. No, you cannot “work any day you want to”. Ask the day laborers that hang out at the Home Depot near my house how they are doing. Some days you pick the wrong parking lot. Some days it’s raining. Some days there just isn’t anyone with a job. The frictions and costs of working day-to-day are huge.
  5. Personally, I think that the following characteristics are associated with “good jobs”. (A) Security/steadiness. As per #2, knowing that your job will be there next week/month/year is incredibly valuable. It allows you to undertake long-run commitments, like marriage, home-ownership, and schooling. (B) Family flexibility. You can deal with your life (i.e. all the crap you need to get your kids to) without the fear of being fired for it. (C) Pay/Benefits. Enough money to afford decent health insurance, or decent health insurance provided by the employer. In short, I think people want stability more than anything. The attraction of those mid-20th century union jobs for workers was that they had lock-it-down certainty about the future.
  6. Yes, it is possible to make any kind of job a “good job”. I used the Costco/Wal-mart distinction as an example. Justin Wolfers and Jan Zilinsky just posted a piece containing further examples. In short, worker productivity is not a fixed value, and paying higher wages is associated with getting higher productivity from the same workers. Costco has a wage/benefit structure that encourages their workers to be productive. In return, Costco saves money from lower turnover. What Wolfers and Zilinsky show is that this works in a variety of settings.
  7. The original post made it sound as if unions were the only way to generate the conditions of “good jobs”. That is not true, and not what I intended to say. Unions were one way to elicit those good conditions from employers, and manufacturing workers were particularly well placed to unionize and negotiate those conditions. But unions aren’t necessary for this. Costco isn’t unionized. [CORRECTION: About 15,000 Costco workers are part of the Teamsters. Roughly 174,000 total Costco workers. DV 1/20/15] We need companies to recognize the value of becoming a “good job” employer, but there are lots of ways to do that.

28 thoughts on “Technology and “Good Jobs”

  1. With regard to your point 3, in the past there was some link of the unskilled (or low skilled) to the owners of capital, of machines, raw materials, etc. Those owners had a strong need for the unskilled, to work with and for their machines, to process and transport their raw materials, etc.

    So, some unskilled were getting substantial amounts of finished goods, of raw materials, of food; wealth, money, from the owners of capital and raw materials, that they could then trade with other unskilled for services.

    But if you get to an economy where the owners of capital and raw materials have little or no need for the unskilled, because of revolutionary robots and computers, (and this is unlike the past you cite), then the unskilled will essentially be in a closed economy. And in this closed economy they can trade, or exchange with the help of currency, all of the services they want with each other, but there will be no food or other goods to trade it for.

    Maybe a simple model will help you to see my view:

    There is an island. One person owns all of the machines, robots, food and other raw materials, power generation, and any other capital. And there are also 10,000 other people on the island with little or no skills, and zero capital or wealth.

    The robots and computers of the single capital owner are so advanced and efficient that they can do almost anything that an unskilled person can do, as well, or better, much better, at a cost in raw materials and goods far below what it would take for a human to subsist. As a result, the island’s single rich man optimizes with the purchase of only the labor of two unskilled humans, paid a subsistence-only amount of goods.

    Now, those two unskilled workers could trade their goods with the other 9,998 for services, but they would die, and so would the other 9,998 as it would obviously be much too little.

    So, perhaps you could tell me how could the 99.98% of workers get enough goods to survive, no matter how many services they bought and sold from each other?

    • Yes, any group of people that is endowed with exactly 0 natural resources and is assumed to be unable to trade at all would starve. The question is why I would think that your scenario represents anything remotely approaching reality. While the richest 1% control a lot of assets, they don’t control all of them. If they were to fence themselves off in Wyoming with their robots, do you think the rest of the US would wither away and starve?

      • I’m not at all sure that it’s that far from reality. Unskilled people have very very little wealth, for many (most?) zero or negative (Later I’ll try to google some numbers, but globally, something like the richest 50 people have as much wealth as half the worlds population, 3.5 billion people.) What the unskilled always had is the value of their labor, which in the past has been very substantial, at least in the first world. If that labor suddenly plunges in value, because its demand plunges due to brilliant machines, then the unskilled, and low skilled, (the majority in the US) will be in serious trouble, that only government will be able to deal with. And, of course, again, its huge to decrease the amount of unskilled with education, and increase the amount of skilled (which also increases demand for the remaining unskilled).

        Before a generation of right wing dominance, we seemed to understand what Nobel Prize winning economist Jan Tinbergen said, Inequality is a race between education and technology.

      • Before a generation of right wing dominance, we seemed to understand what Nobel Prize winning economist Jan Tinbergen said, Inequality is a race between education and technology.

        [ Where is the reference to this quote? The quote is used at times, but precisely where does it come from so we can look to the context. ]

      • According to this census document:

        Click to access Wealth%20distribution%202000%20to%202011.pdf

        (Figure 1) 40% of Americans had a net worth less than $8,000 in 2011. 20% had negative net worth. 60% had a net worth of less than $70,000.

        Even with a large increase in the average investment return (and many of these people are going to have poor, or extremely poor, knowledge of how to invest, and personal finance practices), this won’t last very long if the value of their labor plummets due to a revolution in smart machine capability. And this is especially true with a substantial minimum wage, which for a large percentage of the population may mean a sudden fall off a cliff to zero income.

        I’m not at all sanguine about this, and certainly there’s nothing in theory that says this kind of disaster can’t befall a very large percentage of the unskilled, if not most of them.

      • Richard – not ignoring your comments. But I want to write a longer post about this issue, so I’m holding off until then.

      • Thanks Dietz,

        I’ll add a few other thoughts. You might say that if capital, including super AI machines, takes off, then the returns on investments will take off.

        This is an interesting question. It depends on how much market power firms have, and how quickly competition can dissipate the cost-plunging and quality-taking-off to the consumers, in dropping prices. My initial guess is that once the revolution really gets rolling, for 10, 20, 30 years returns on stocks would be very high, even unprecedented, but long run competition will return them to about normal. Although long run things could be just completely different in the world. These robots/computers could completely transform the world in the end game, doing virtually everything, so almost no one can work, and creating incredible total wealth. It would be a completely different kind of economy and world.

        In any case, in a world where most of the unskilled and low skilled have close to zero wealth from their lifetime labor, you could say, let’s give them some robots! Let’s give everybody some stocks to live off of.

        Well, think about that. In the past a low-skilled person might have lifetime labor value worth, say, $1.5 million in present-value. So, let’s just now give them $1.5 million in stock in an index fund!

        But, I have a career in personal finance. If someone has $1.5 million in present-value of lifetime labor they can’t blow that (aside from becoming an alcoholic, or something like that). Some shyster can’t just sucker it away from them over the next year. They can’t decide to spend it all on a giant house and a Mercedes (and even with credit, there’s bankruptcy, and a limit of 25% in wage garnishment, zero in some states).

        You’re for the most part forced to spend your lifetime labor slowly and steadily. And, sure you could lose it with disability or old age, but so far, no amount of billionaire powered propaganda has gotten people to vote for not providing at least some minimum for the aged and seriously disabled.

        But give people a lump of liquid wealth to replace the plummeting in the value of their lifetime labor, and that they can blow easily. Even if they’re very responsible, shysters everywhere (and even well intentioned people) can get them to invest in a restaurant, some “hot” single stock, commodities, derivatives, sector funds, actively managed funds that are far riskier than they understand, an invention, and so on.

        A close example we see today, where a very large percentage of lifetime wealth is liquid, is with retired athletes, and we see disastrous results. According to this Sports Illustrated article in 2009:

        Although salaries have risen steadily during the last three decades, reports from a host of sources (athletes, players’ associations, agents and financial advisers) indicate that:

        • By the time they have been retired for two years, 78% of former NFL players have gone bankrupt or are under financial stress because of joblessness or divorce.

        • Within five years of retirement, an estimated 60% of former NBA players are broke.


        So there’s something that’s just very different, and a lot safer for most people, in having the bulk of your wealth in labor endowment instead of something relatively liquid.

        Of course, the government could instead provide free high quality health care, college, preschool for children, and other Heckman-style early development, abundant public recreation, schoolfare,…, but with the billionaire-powered rights ability to block in our system, it could get very ugly and tragic before any of that happens.

    • Also, these net wealth figures can be predominantly housing, and you can’t eat your house, or get electricity or gasoline from it, and so on. If the value of your labor drops below minimum wage, most people won’t last long without government help, and we have one major party, with great power to block in our system, that’s happy letting large numbers move to homelessness.

  2. Or, we just use technology to reduce the costs and frictions of the employment market. Like Uber and Airbnb, only for labor sharing. Also, see oDesk, a technology platform for freelance labor.

    At least around here a large part of the “friction” associated with the day laborer market is payroll taxes and immigration (i.e.. avoiding both). $100 bucks a day will get you about 8-10 hours of work, with no questions and no taxes, and no surprise safety inspections. Even if they are employed 80% of the time, they are getting cash, no questions asked, and effectively making more than min wage when you factor in costs of getting caught. Markets tend to gravitate to a natural equilibrium with a central clearing house. If the employers and employees had their druthers, they would go to the same place every day. Underground markets move around to avoid getting caught.

    • Even if technology makes the day-labor market really efficient (so I don’t have to worry about picking the right parking lot), that doesn’t solve the underlying uncertainty problem. If you married that technology with some kind of income-smoothing insurance program, then you might have something.

      • Except that UI only kicks in under certain conditions. If you work on a day to day basis, you do not qualify for UI.

      • “Doesn’t solve the underlying uncertainty problem”

        So suppose we had “universal” UI and an app for labor sharing. Does that really “solve” the uncertainty problem? Even if labor demand is “uncertain” why can’t everyone just adjust their wage bid/offers accordingly (If I were king the market would clear based on a dutch auction)?

        Let’s put this somewhat differently: Suppose in 2139 labor markets were efficient, all jobs were “good,” would we still see involuntary unemployment in 2139? How can that be if all jobs are “good”? If the biggest decision in 2139 is whether to postpone my year-long trip to Rome because I could not work enough, do I really care about “involuntary” unemployment?

      • If you had an app that could clear the labor market “instantly”, then “involuntary unemployment” would go way down because people would be able to search and match more quickly.

        But I don’t think it would disappear. Why? (1) I think firms use workers as a lumpy input. That is, there are fixed costs for firms to bring someone on, train them, etc.. that generate frictions. So it is quite possible that firms won’t want to hire people no matter how low the wage is. (2) People are not infinitely mobile. The only way to clear the market may be to move 1,000s of people from Baltimore to Chicago. That is not something that can happen instantly.

      • “If you had an app that could clear the labor market “instantly”, then “involuntary unemployment” would go way down because people would be able to search and match more quickly. ”

        What I was getting at is that there are other sticky prices in the economy, and how much involuntary employment there there will be will be a function of other sticky prices, not merely labor. Robots will be making a lot of things, but will all markets for all things clear simultaneously? Debt itself locks in the price of a good for years. The biggest, stickiest price in the consumption bucket today: housing, where prices in the US are locked in for 30 years. In a world of robots serving most of our needs, how are these tourist attractions financed? If the price of capital is locked in today, and the capital delivers a fixed quantity of services over multiple periods, P and Q are fixed so any demand fluctuations will result in periods of excess demand or supply of those services, and we ought to see involuntary unemployment.

        I will have robots to make food, robots to make housing, robots to make robots. what if I have too many food robots because there is a new fad diet?

        …You can have a world filled with robots that make essential goods for all of humanity, yet still have involuntary unemployment, because the market for residual labor does not clear simultaneous to the market for trips to Rome.

      • Got it, i didnt see your point first time around. But i agree with your story here – there are always going to be coordination failures bc no one has full information.

    • Good catch. But I can’t find support for the 20% number. I found 15,000 workers represented by Teamsters, out of total of 174,000 employees, so 8.6%. If you have a different source, let me know.


    January 19, 2015

    Technology and “Good Jobs”

    Before a generation of right wing dominance, we seemed to understand what Nobel Prize winning economist Jan Tinbergen said, Inequality is a race between education and technology.

    — Richard Serlin

    [ A day later, I am still waiting for the precise reference to the Jan Tinbergan quote. Where precisely is the quote located, and what is the context? ]

      • Click to access On%20Growth%20and%20Income%20distribution-final%20version.pdf


        As Jan Tinbergen (1975) pointed out, observed relative wages are the outcome of a “race” between the forces increasing the supply of skills – mainly education, but also experience – and those increasing the demand for skills required by firms – technological change. Tinbergen stated that “in the past seventy years that race was won by education” (1975, p. 103) thus resulting in falling
        relative wages….

        [ Terrific, I will go to the original source tomorrow. ]

      • I am grateful, now to look to the context of the paraphrase from Tinbergen and ask whether the paraphrase is applicable today, or where and why is it applicable or not.

        The Tinbergen paraphrase strikes me as used as a way of excusing structural-social limitations making for inequality. I do not think education is enough of an answer to inequality in America today, but I need to think this through.

  4. Let’s not forget the generations of thinkers who have been worried about the alienation from the output of one’s work that is inherent in production line manufacturing.

    The most common highlighted passage in e-versions of Gladwell’s Outliers: “Autonomy, complexity, and a connection between effort and reward—are, most people agree, the three qualities that work has to have if it is to be satisfying.”

    It’s plausible that these qualities are more likely to be found in many of the service sector jobs you listed than in large scale manufacturing.

    • I’d agree – at a minimum, there is nothing that makes manufacturing work score any better on those three characteristics.

  5. Clear and persuasive. I want to point out one additional thing that is illustrated by your list of examples of service jobs: those with lower education requirements are not necessarily more susceptible to automation. Over the last few decades, lots of engineering person-hours spent with slide rules have been replaced by CAD/CAE software. How many florist hours have been eliminated by robot flower arrangers? Of course, we still have lots of engineers, because calculation is only part of the job. Automation will continue to be extended, and there will be painful adjustments (which we can mitigate by policy choices), but I don’t see any evidence that we will not have jobs for all skill levels for the foreseeable future.

  6. I read the linked paper but still very much disagree with the concept that you can make any job into a good job as defined above.

    Of course an employer can pursue different paths. In path A you can hire smart, skilled, motivated, honest workers and support them with capital and technology to optimize their productivity and you can pay them well with stability and benefits. In path B you can take just about anybody and get them to do some simple job at a low wage until such a time as they leave to go to a better job. Obviously there is a limitless range between the two extremes and most firms actually have a mixture of A and B.

    The success of the firm is based in part at how well they manage this balance, with the caveat that there is may be competing paths to success, and the balance is dynamic and often influenced by the pats taken by competitors

    I agree you can TRY to make any or every job into A. And I suppose it is always possible to do so. However it may not be most efficient to do so. Model B may still be the most economical way to produce a widget in that given market for that firm.

    Even more importantly though, this entire debate assumes everyone is capable of being skilled, smart, reliable, honest and so forth. This is absolutely not the case. It isn’t true in developing nations (they are not necessarily skilled or educated), nor is it true of everyone in the developed nations. Not everyone is above average, and by re-engineering jobs from B to A, these people are put at a serious disadvantage. Indeed they become superfluous to the extent that they cannot be molded into a type A. And not everyone is or can be.

    All this said, I agree that there is value and positive externalities in the evolution from type B to A. No doubt. But the process cannot be forced without negative consequences especially to those individuals least able to adopt to type A.

  7. Pingback: When an Op-Ed About Growth Fails | The Growth Economics Blog

Leave a Reply

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

You are commenting using your 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 )

Connecting to %s