A Social Safety Net for the Job Apocalypse

Summary: As previous posts have described, the next wave of automation will the big one. We need to begin preparing for it now. Economist Ed Dolan describes one important part of the policy mix we will desperately need to manage this aspect of the coming new industrial revolution.

Businessman kicked by his boss - a foot on spring.
By Sergey Nivens at Adobe iStock – 98043068.


A Social Safety Net for the Job Apocalypse

By Ed Dolan.

Automation and artificial intelligence are increasingly disrupting the labor market. Some see a coming “job apocalypse.” Others simply see a continuation of existing trends toward growing inequality as more and more people fall out of the middle class into less stable, lower-income employment. In either case, we are not prepared.

One of the many things that need attention is our social safety net, especially those parts of it that deliver health care and income support to those in need. This post suggests two major safety net reforms that would mitigate the impact of coming shocks to the labor market.

A scenario.

A specific scenario will help focus the discussion – one based on a job category that is likely to grow in the coming years even as many routine and repetitive jobs disappear. That category comprises home health aides, personal care aides, and other occupations that can be loosely referred to as “eldercare.” These jobs are resistant to automation because they are neither repetitive nor routine. Moreover, they are jobs where people value the human touch.

So here’s the scenario: You have a nice job as a welder, which, for an experienced worker, pays about the median family income. Out of the blue, you are replaced by a robot. After a while, with your unemployment benefits running out and a growing conviction that you are never going to get another production job, an employment advisor suggests you try work as a home health aide.

There is plenty of work to be done. You have your choice of working for an agency that will place you with clients in need, or you can become self-employed, taking advantage of on-line platforms that match caregivers with customers. The catch is, you won’t be able to earn much more than $20,000 a year as you start out, and even as you gain experience, you are never going to get back to what you earned as a welder. If you’re trying to support a spouse and two kids, you’re not even going to make it up to the poverty level. Plus – your benefits are gone, your work schedule is unpredictable, and your back hurts more from lifting patients in and out of bed than it was even when you were zapping big chunks of steel together.

What reforms to the social safety net could help you survive in this new world of work?

Priority one:
Decouple health care from employment.

The first thing you are going to need is dependable access to health care for yourself and your family. You may miss the kind of employer-sponsored health insurance (ESI) you used to have as a welder, but although most people who have ESI say they like it, as public policy, it has serious flaws. Three of them will become even more troublesome if a full-fledged job apocalypse strikes:

  • Job lock: Workers fear to change jobs, to go into business themselves, or to transition to the gig economy because they fear that they may lose access to health care.
  • Inequity: Employees in the top fifth of the income distribution get an average of $4,500 in benefits from ESI while those in the bottom fifth average just $500 in benefits.
  • Fragmentation: Thousands of ESI plans mean fragmentation of the health payment system, driving up administrative costs and opening new avenues for rent-seeking by providers.

How can we ensure that people without employer-sponsored insurance don’t lose access to health care altogether? Some would-be reformers advocate a “single-payer” plan like Medicare for All, but there are legitimate doubts about the affordability of that approach, and many who currently have some kind of coverage fear being forced into a one-size-fits-all government insurance plan. There are better ways to fix our healthcare system. Here are three options that could be fully funded even without increasing what the government already spends on health care:

  • My favorite: Universal Catastrophic Coverage. UCC would fully protect everyone from financially ruinous medical expenses while asking those who can afford it to pay a fair share of the cost of their own care.
  • A promising GOP alternative: Bruce Westerman’s Fair Care Act (FCA) would use government-sponsored reinsurance to make private coverage universal and affordable.
  • A Democratic Band-Aid: Several presidential candidates propose adding a public option to the ACA {aka ObamaCare}. That wouldn’t go as far as UCC or the FCA, but it would give coverage-of-last-resort to people who lost, or could not afford, employer-sponsored coverage.

Priority two:
A floor on income that leaves work incentives intact.

Low pay is likely to be an ongoing reality for home care aides and many other service workers. If that is true, then the only realistic way to avoid ever-increasing inequality is to give people a guaranteed minimum that is high enough, when combined with the income from modestly paid work, to allow a decent standard of living. The guaranteed minimum would also give something to fall back on in times of crisis caused by recession, family obligations, illness, or disability.

Many people worry that offering people a guaranteed minimum, even at a sub-poverty level, would spawn packs of loafers living in pup tents on the sidewalk and spending their handouts on drugs and alcohol. Like most economists, I do not share that view.

The influence of income on work effort has been widely studied. Some studies are based on how people react to lottery winnings, inheritance, tribal casino dividends, and the like. Other studies use structural data on wages, salaries, and work behavior. Typically, such studies do find that, other things being equal, people tend to take work less and take more leisure as their incomes increase, but the effect is small. The majority of studies of “income elasticities,” as they are called in economic jargon, indicate that other things being equal, a 10% change in income will induce a zero to 1% increase in hours worked. And that does not mean that a 10% boost in income would cause 1% of beneficiaries to stop working altogether. More often, it would mean continuing to work but taking a vacation or day off now and then.

In real life, though, while income does tend to have a small negative effect on work effort, the amount of take-home pay a person gets from an extra hour’s work has a much stronger positive effect. Isolating that effect, it appears that a 10% increase in take-home pay tends to cause a 1 to 4% increase in hours worked. Economists call that the “substitution effect.”

A properly structure reform of the social safety net can take advantage of the fact that the strong substitution effect of higher take-home pay outweighs the weak effect of higher income. Here is what needs to be done:

First, as I have explained in detail elsewhere, any guaranteed minimum income or basic income that is used to provide a floor to people’s standard of living must replace our current fragmented, in-kind system of means-tested welfare, not be added on top of it. The problem with the current system is the way it aggressively cuts benefits for each dollar earned. A study from the Congressional Budget Office shows that when the effects of the benefit reductions in EITC, SNAP, TANF, CHIP, and Medicaid are combined, a family with income just above the poverty level may keep as little as 20 cents for each added dollar of earned income. A guaranteed minimum income would let low-income workers keep all, or nearly all, of their added earnings, greatly increasing work incentives.

Second, if you think that is not enough work incentive, the reformed safety net could pay out part of its benefits in the form of wage subsidies for low-income workers. The earned income tax credit already does that to a limited extent, but its incentive effects could be considerably strengthened. One change would be to pay benefits in the form of extra cash added to the monthly paychecks of low-paid workers, rather than as an annual tax credit. Another would be to extend the income range over which wage subsidies operate. A third would be to allow people without children to qualify for the program.

A guaranteed minimum income plus wage subsidies for low-paid workers is a combination that I call Integrated Cash Assistance. You can find a full description here.

A final caveat

Let me end with a big caveat: No reform of the social safety net is going to cure all our problems. The shock of accelerating changes in the labor market, whether we call it a job apocalypse or not, is going to disrupt many lives. The effects will go far beyond the merely economic. People whose lives are upended, even with the best safety net, are going to be vulnerable to stress, depression, and anxieties that will in some cases have ugly outcomes like substance abuse, suicide, domestic violence, and child neglect.

We should not expect cash assistance alone to cure all these ills. More narrowly tailored, hands-on help will continue to be needed in many cases.

Still, I believe that a reform of the social safety net that offered universal access to health care, decoupled from employment, and replaced our current system of in-kind, means-tested welfare with a program of integrated cash assistance would make America much better prepared for a job apocalypse.


Editor’s afterword

Dolan’s analysis is typically clear and insightful, describing effective palliatives for widespread job losses. But these probably would occur in a nation of inequality rising beyond today’s already record-high levels. I wonder if middle-class workers downgraded to low-skill jobs or welfare would passively accept their fate – or try their luck at revolution (either communist, fascist, or other).

See my posts about the GMI, an effective palliative (but nothing more).

Other posts in this series

Ed Dolan

About the author

Edwin G. Dolan is a Senior Fellow at Niskanen Center (Ph.D. from Yale). He was an Asst. Prof. of Economics at Dartmouth, and later on the faculties of U of Chicago, and George Mason U. From 1990 to 2001, he taught in Moscow, Russia, where he and his wife founded the American Institute of Business and Economics (AIBEc), an independent, not-for-profit MBA program. He has taught at several universities in Europe, including Central European University in Budapest, the University of Economics in Prague, and the Stockholm School of Economics in Riga. He is a Senior Fellow at The Niskanen Center.

During breaks in his teaching career, he worked in Washington, D.C. as an economist for the Antitrust Division of the Department of Justice and as a regulatory analyst for the Interstate Commerce Commission, and later served a stint in Almaty as an adviser to the National Bank of Kazakhstan. When not lecturing abroad, he makes his home in a small town in northwest Michigan.

His publications include Introduction to Economics (2014), TANSTAAFL, the economic strategy for environmental crisis (1971), and The Foundations of Modern Austrian Economics (1976). See his articles at Roubini’s Economonitor, Business Insider, the Niskanen Center, and his blog.

See his twitter feed. and his other posts at the FM website.

For More Information

Ideas! For shopping ideas see my recommended books and films at Amazon. Also, see Chapter One of a story about our future: “Ultra Violence: Tales from Venus.

If you liked this post, like us on Facebook and follow us on Twitter. For more information, see these posts about employment, about the coming wave of automation, and especially these …

  1. The promise and peril of automation.
  2. At last, economists see the robot revolution. Here’s why they worry.
  3. The robots are coming, bringing hope of a better future.
  4. A warning about the robot revolution from a great economist.
  5. How Robots & Algorithms Are Taking Over.
  6. Tech creates a social revolution with unthinkable impacts that we prefer not to see — About sexbots.
  7. AI will reshape the world. Films show how.
  8. The new industrial revolution will change everything.

For deeper analysis see these books…

The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies
Available at Amazon.
The Rise of the Robots
Rise of the Robots: Technology and the Threat of a Jobless Future

24 thoughts on “A Social Safety Net for the Job Apocalypse”

  1. The UK has been doing the first of these for a couple of generations – health care decoupled from employment. Care which is state financed, universal, and free at the point of use. Or largely free – there are flat rate prescription charges for those under 65. The UK does this through a nationalized health care system. In Continental Europe you have a somewhat similar result from a Bismark-type insurance model.

    The UK system has advantages and disadvantages, but it does mean that no-one loses access to health care, no matter how expensive their treatment need, as a result of changes in employment.

    The UK is also trying at the second. It has also had a minimum wage system for quite a while now – I think for well over a decade – and is attempting, with very mixed success, to integrate this with the benefits system in a way that makes a transition from benefits to work economically attractive. It has turned out to be a very thorny problem, but they have been trying to make it happen also for about a decade.

    Their success in the second is qualified not only by the intrinsic difficulty of the task, but also by the environment in which they are trying to do it. Its one in which housing costs are going through the root, at least partially because of increased demand from substantial immigration coupled with considerable restrictions on increasing the supply of housing.

    Whether this is a response to a situation caused by automation I am not at all clear.

    1. Henrik,

      “The UK has been doing the first of these for a couple of generations – health care decoupled from employment.”

      Every developed nation other than the US has done so for a long time.

      “It has also had a minimum wage system for quite a while now”

      That’s not a minimum income system.

      1. “That’s not a minimum income system.”

        No, a minimum wage is not a minimum income system.

        But, as I said, they are trying to combine it with the benefits system in such a way as to both to give a minimum income and to incentivize moving off benefits into full time work. The two together amount to a minimum income system.

        The benefits reform is called Universal Credit and has been very difficult to roll out, and is quite controversial. It tried to combine all the previous fragmented different benefits into one payment with provisions which make it attractive to get back into part time and then full time work.

        I am not familiar with all the detail. I do think the aim of the combination, is to deliver minimum income, and that is the outcome. But its to find a practical way of doing that which includes paid work.

        I am not sure to what extent the reforms are targeted at situations which are the consequence of automation. The UK has relatively little automation compared to other EU countries, partly because employers have been able to rely on low wage labour from the recent joiners of the EU. With Brexit this is vanishing as the immigration proposals post Brexit are basically to eliminate unskilled immigration. The connexion between immigration and minimum income issues is important and overlooked.

        About the health care system. The UK is different from almost all other countries in that it is not an insurance based model. It is not a pay-reimburse system. There are no contributions to care, other than the flat prescription charges which are waived for benefit recipients. In the European countries where I have lived and worked, become unemployed and you will have financial problems with health care.

        Not as acute as in the US, but you will have them. In the UK you will simply not notice any difference.

        If you want a practical illustration of this, there is currently talk of Northern Ireland joining the Republic, the so-called ‘united Ireland’ proposal. This throws up many changes which the current British citizens of Northern Ireland would encounter, and one of them is health care.

  2. I would use Dolan’s approach and assumptions to power other needed changes.

    Use the the coming AI explosion to develop medical AI. Let’s call the AI. Dr. Robot. The goal is to drive down costs while providing more coverage. Even though Dr. Robot will likely be at the doctor’s side at first, the potential for having AI in a triage will be evolutionary, due to the expected exponential increase in AI ability and numbers.
    Use the coming AI explosion not to just help the welders, etc., but also the mangers, professors, and other intelligent and competitive persons to become physicians or certified nurse practitioners. This will also help drive costs of the system down.
    Use AI to regulate and eliminate planned obsolescence in order to make repair which “are neither repetitive nor routine. Moreover, they are jobs where people value the human touch.” An example is the lifespan and perceived worth of guitars. With the growing numbers of guitars and their longevity, repair will be a growing sector.

    I would add one other criteria for the scenario: The need to bring economic parity to world before the AI revolution makes economic investment for lifting world poverty moot. The assumption is that the AI revolution is a large enough problem that having a large fraction of humanity with the likelihood of devolving into smash and grab military cultures is counterproductive. Otherwise, my opinion, “War is Peace” and other horrors await us.


  3. Another solid article, Larry. I know that this one was written by somebody else, but finding and vetting such articles isn’t easy, I know this because I’ve done it with mixed success.

    1. Pluto,

      Thank you for the kind words. The sad aspect is that articles like this and yesterday’s about Coronavirus are what people need, but not what they read. Americans want existing misinformation and fake news! So that’s what we get. Free markets give what we want, not what we need.

      The pitiful aspect of this is that we then whine about the quality of the news we demand!

    2. I take credit/blame for the content of the article, but Larry deserves a lot of credit for good editing and a lot of thoughtful input on earlier drafts.

  4. Dolan-

    Excellent article. Only nitpick is that in the short/near term, the tradesmen (welders, plumbers, and electricians) are going to do well. I’d submit it is corporate middle management that is most at risk. And, the latter group has large college tuition debt to contend with.


    1. Mike,

      Thanks for the review!

      Remember, this is an article about macroeconomics — ie, about aggregates. Not granular, looking at the effect of automation on specific jobs. Dolan looks at one kind of job as an example.

      So there will of course – as always – be winners and losers from automation, the aggregate (aka net) effect will be to increase unemployment. The number of tradesman jobs probably will change little. The demand for servants by the rich might increase irrespective of automation (human servants probably being a status item). Etc, but not enough to offset the effect of automation reducing jobs.

      That does not mean exterminating broad classes of jobs. For most it will mean automating kinds of tasks, so that people in a specific job can be more productive – and so fewer are needed. Think of automation affecting radiologists: machines helping them do their job faster and better. That probably does not increase the demand sufficient to offset the number of job hours lost.

    2. Mike: “I’d submit it is corporate middle management that is most at risk.”

      Reasonable but I wouldn’t take that prediction to the bank. There isn’t enough data to make this more than a guess, though. These things have a way of twisting in odd directions.

    3. Mike– Yes, we should not underestimate the exposure of middle management, not to mention professionals like lawyers, accountants, and bond traders to Automation/AI. With regard to welders–it will matter if your welding job is nonrepetitive (eg patching together complex broken equipment) or repetitive. Humans standing on an assembly line welding parts of a car together are already a thing of the past.

      1. Ed and Mike,

        I suspect that many middle manager jobs will vaporized – for two reasons.

        First, the jobs that they supervise will go away. How many people are in effect clerks? They fill out paper, move paper, input paper, file paper, and retrieve paper? Much of that is going away as the entire processing process goes electronic. When those workers go away, so will their managers.

        Second, a lot of management will go electronic. System track people, so that much of the subjective employee assessment will become objective numbers. Think of the credit approval process. There used to be a relatively senior credit officer in every branch bank – interviewing people, reviewing their applications and updates, etc. Now an algo does all of that both better and cheaper. Managers will still be needed to manage people, but they’ll be more junior than today – and managing more people – because systems will do much of today’s “management work.” See call centers today, where one manager handles a team of 12 to 18 people.

        When I was an internal consultant for Schwabl’s president in 1989, I showed him (to his surprise) that they had more management layers than did the Roman Catholic Church (for an organization with size and geographic scope far smaller). They had an average span of control of (from memory) 4 per manager. Now the average for organizations is probably closer to 7 (speaking of “white collar” jobs). It will grow more, imo. And so many layers will disappear.

        Third, better automatic information processing will eliminate the need for so many layers. Managers will see Dashboards with operating metrics, with data that formerly was assembled, collated, and analyzed by 2 or 3 layers. For example, one meteorologist with a modern system can monitor storms (rain/snow, winds – up to hurricane intensity) with no subordinates, making better decisions that used to require a dozen experts.

        The changes we see today that seem so large are nothing compared to what’s coming.

      2. Larry / Ed-

        To Larry’s point, the leadership needed to lead in middle management is vastly different than what exist today. It’s more of decentralized, complex decision making and being able to find patterns across large swaths of data. The people factor will decrease by a significant amount. I think that will hit us the hardest.

        For tradesmen, outside the factory, many tasks that plumbers, electricians, and welders perform ARE simple and repetitive. However, terrain and environment constrain the ability to get a robot in. So, I think the tradesmen are safer for now.

        On the macro level, in the long run, it will be completely different.


      3. Mike,

        “the leadership needed to lead in middle management is vastly different than what exist today.”

        I frequently hear this in discussions of automation – the jobs will become increasingly complex and difficult, the opposite of disappearing. It’s not been true so far, and will increasingly not true because it misunderstands what is happening.

        Those jobs are disappearing as organizations flatten and algos replace much decision-making (as described in my previous comment). You describe the fewer jobs that remain after the next wave of automation: decentralized, complex, analytical (or the opposite, requiring people-skills). That’s why those jobs will remain, as they are what happens to a layer of management after that layer and the layer below have simpler functions automated (ie. lots of jobs lost).

      4. One more note-

        Back to the assembly line. I used to work in tech in manufacturing. While the roles changed, they did not go away. Instead, work used to be siloed. Now, R&D folks have to work side-by-side manufacturing and supply. Tesla is a good example of this. Positions will decrease over time, but they will not go away. It just evolves.

      5. Mike,

        “Positions will decrease over time, but they will not go away. It just evolves.”

        That’s a rebuttal to the “they’ll ALL GO AWAY” straw man. Nobody says that they’ll ALL GO AWAY. There will just be fewer jobs. Hence more unemployment.

  5. The AI revolution provides us with some challenges. One example: how do we provide opportunities for people, especially young men, to make a positive contribution to society and to have a legitimate purpose for their lives? Failure to so will invite violence and revolution invited by people promising opportunity and purpose to the disenfranchised. Creating a society full of people with little to do and serious purpose in their lives is like letting deadwood accumulate in a forest and wondering why the wildfires are so bad.

  6. This eminently sensible analysis has close echoes with arguments/ideas put forth by Charles Murray (written In believe in 2006) entitled “In Our Hands: A Plan to replace the Welfare State” where Murray proposes the elimination of all dimensions of welfare (food stamps etc.) in exchange for a basic income of $10.000.00 per year (functioning as a type of subsidy)–in the hopes of strengthening family structure rather than the ever expanding role of the administrative state.

    But Larry, I believe you have raised an even more fundamental question when you state “I wonder if middle class workers downgraded to low skill jobs or welfare would passively accept their fate or try their luck at revolutions( either communist, fascist or other).”

    We may now be in a position where our constitutional state is now being overwhelmed by the administrative state as well as the national security state–in a context in which feelings of individual powerlessness, worthlessness and generalized senses of humiliation are accelerating.

    Is simply seeing the world more clearly– enough–in such a circumstances?

    Don’t future policy proposals for reform have to somehow incorporate within them an answer to a growing sense of the increasing meaninglessness of modern-day life?

    Is this were issues of belief/theology/faith and their role in motivation(overcoming passivity) and finding meaning–should come more and more into play as well as in-depth discussion?.

    1. Pluto,

      That is a great – and sad – book!

      Of course, they had an excuse for their screwed up world – with 30% of the population dead.

      The TV series “The Last Ship” took a light look at a post-plague world. It probably would be far worse than anything in that show.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Scroll to Top
%d bloggers like this: