More and more jobs are vanishing, and they aren’t going to come back.

But it’s a weird sort of vanishing. Until the late ’70s, our increasing worker productivity meant more pay per worker, and a drop in prices of goods. At that point, we started to need fewer workers to make all the stuff anyone would want, and that meant that wages dropped, too. Some of this briefly got hidden by “offshoring” — it was cheaper to move manufacturing to China, then to Bangladesh, and so on — but now those countries are starting to see automation take jobs away, too. The cost of production is dropping to zero.

What’s really stupid about this is that it leads to people starving amid plenty. That’s because we actually use jobs for three things:

  1. To make things we need.
  2. To allocate resources (via things like wages).
  3. To achieve individual meaning in our lives.

The basic problem is that we can achieve goal #1 with capital equipment right now, and since our companies are shareholder-owned rather than worker-owned, that means that they do so quite cheaply, but in the process stop satisfying goals #2 and #3.

And so you end up with a bunch of people un- or underemployed, frantically struggling to make ends meet, even as productivity is at an all-time high. There are lots of goods, but the allocation of resources becomes intensely lopsided. And there are lots of reasons for people with goods to want it to be lopsided; Peter Whoriskey’s Washington Post article “‘A way of monetizing poor people’: How private equity firms make money offering loans to cash-strapped Americans” illustrates one of many.

Basically, people with limited economic options are a resource to be harvested, if you have a mind to. If you want to go deeper, here’s an article I wrote a while ago explaining in more detail why “a pool of poor people” is really valuable if you’re rich — as well as what “rich” actually means.

But using poor people as an economic resource is valuable only to a small fraction of society. To most, the idea of “everyone starves to death” seems pretty unpleasant, not to mention idiotically unnecessary if everyone is starving to death because we have too much wealth.

If you have a decent job right now, how much of a raise would you want to accept having most of your job responsibilities switched to shoveling feces?
… But then, why do people whose actual job is to shovel feces get paid so much less than you do?

A universal basic income proposes to address these issues by basically paying out a dividend to everyone from the overall wealth society is gaining. This has been done in some limited cases (in Alaska and Saudi Arabia from oil wealth, along with various local experiments) and has worked out not badly. However, there are three important critiques of universal basic income that I think are worth paying attention to.


Critique #1: People Won’t Want to Work

The first is that “people won’t want to work.” This is sort of true: People won’t want to work at shit jobs, unless you pay them a lot to do it. Which is kind of obvious. I mean, if you have a decent job right now, how much of a raise would you want to accept having most of your job responsibilities switched to shoveling feces? I’m guessing you’d want a not-small one.

But then, why do people whose actual job is to shovel feces get paid so much less than you do?

The answer is simple: They get paid less because the only people taking that job are people who have no choice. You can make feces-shoveling a low-paying job if, and only if, people can be forced into it. And a universal basic income would break that.

The reason it breaks that is all about this curve, which roughly plots how good your life is as a function of how much money you have.

Basically, whenever something goes wrong in your life, how serious it is depends on whether you have enough money to cover dealing with it. For some things, no amount of money will help — if your spouse is killed in a car crash, if you are hit by a meteor. For other things, even a small amount of money makes a big difference; for example, a flat tire is an inconvenience if you can get it fixed, but a disaster if you can’t afford a cab and this makes you late to your shift job, which means you get fired.

The most important measure of your wealth, in a lot of ways, is how much of a shock it would take to cause the negative consequences to spiral out of control, and how often such a shock is likely to hit. If the cost of a flat tire is enough to cause a major disaster for you, you’re poor; shocks that big happen all the time. If the cost of a cancer diagnosis is catastrophic, you’re in an unstable middle class; that may or may not hit your family. If virtually nothing can destroy you financially, you’re rich. What this boils down to is the curve above. Money doesn’t, in fact, buy happiness: People’s reported life satisfaction levels out once you go above a certain not-too-high wealth threshold. But poverty can buy you one hell of a lot of misery. Specifically, it can put you in situations where you have to do something that will harm you badly in the long term in order to survive the short term.

In terms of the wider economy, someone who has to accept a long-term loss to survive the short term is in a very weak bargaining position — and that sort of weak bargaining position is exactly what breaks the free-market hypothesis that “trade makes everyone richer.”

To explain: If two people engage in free trade, swapping cheese for chalk or whatever, then by definition they both come out happier, because if they didn’t, they wouldn’t have engaged in that trade. This observation goes right to the core of why markets can work so well, and why they increase both the total wealth of society and the individual wealth of people in it.

But that statement has a very important “if” in it: that the trade is free. If two people have radically different negotiating leverage, that “if” ceases to work. One person can demand that the other person give up nearly anything, enriching themselves in the long term at the expense of another person, not to their mutual benefit. Worse, if wealth is a form of power (it always is), this can lead to the negotiating imbalance getting worse over time, creating a feedback loop that ends in absolute monarchy.

A universal basic income breaks this in a fundamental way, as does socialism in any form. Basically, if people have enough access to the necessities of life that they aren’t on the steep left edge of that curve, be that access in the form of cash payments or universal access to housing, health care, and so on, then nobody is going to take a shit job unless they are paid enough to make it worth it.

This would overturn the status hierarchy of our world very deeply. Right now, the hierarchy of jobs by how much wealth they get you is roughly:

  1. Jobs where you can directly tap into existing large concentrations of wealth and take some for yourself.
  2. Jobs which are very supply-limited, so people who can do them are in high demand.
  3. “Ordinary” jobs, not in any other category.
  4. Terrible jobs, which people only take if they have no other choice.

With a universal basic income in play, suddenly a “terrible” job turns into a job limited by finding people willing to do it, and so suddenly you end up with:

  1. Jobs where you can directly tap into existing large concentrations of wealth and take some for yourself.
  2. Jobs which are very supply-limited, either because not many people can do them or not many people are willing to do them, so people who can do them are in high demand.
  3. “Ordinary” jobs, not in any other category.

Basically, you could find sanitation workers getting paid more like professionals. Right now, if you’ve got a job in group #3 or #4, and you want a job in group #2, tough luck: There aren’t many people who can do them, and access to the training is often limited and expensive. Your best chance of getting a group #2 job is to start out with parents with a group #1 or #2 job. But in a post-universal basic income world, there are plenty of jobs that anyone could do in group #2. You want to be paid more? Sure, you just have to take on a nastier or harder job. You can make that trade if you want.

It’s much more meritocratic than the world we’re in right now, and pulling us away from the uneven-bargaining loophole in the free-market hypothesis means that far more exchanges satisfy the rule of “everyone profits.” So one would expect overall societal wealth to go up. (One would also expect overall wealth to go up because things like “people dying of preventable diseases” or “people not being able to work very well because it’s hard to get a good night’s sleep on the streets” are deadweight losses to everyone, decreases in productivity that have no compensating increase.)

So what this shows about the first critique of universal basic income is that it’s really far more complicated. It’s not that people won’t want to work; it’s that wages for jobs which had previously been very low might suddenly become very high, and so there’s going to be a reshuffling of job value.

Don’t underestimate the significance of this. A very robust finding in behavioral psychology is that people are overall okay with changes in wealth and power, unless those changes cause swaps in relative status between people. Suddenly moving from the top to the bottom, or even just down one rung, is a scary business. If we want to implement anything that looks like a universal basic income, we need to think very seriously about how people will deal with that.

Critique #2: We Can’t Afford It

An obvious critique is: “How would we pay for it?” This argument makes a lot of sense if you think about a basic income as a small perturbation of our current politics, but it falters if you zoom out for a moment. If you measure the net wealth of our society, it’s increasing rapidly. This is a graph of the U.S. net domestic product (GDP less capital depreciation, a good if rough measure of total wealth production per year) since 1930, courtesy of the Federal Reserve Bank of St. Louis:

If you divide this out over the U.S. population, you get a per-capita curve that has much the same shape. Here’s per-capita GDP, since I couldn’t quickly get a per-capita NDP plot and I’m feeling lazy:

The current mean net domestic product per household (not per person) is about $129,000; compare that to the current median household income of $59,039, or the mean of $72,641. There is a tremendous amount of wealth gain going on, and we need to recognize that our decision to allocate it via corporate mechanisms which prioritize shareholder revenue — followed by a tax structure which penalizes income significantly more than capital gains or held assets — is a decision, not a law of nature. The money, or more to the point the real wealth — actual goods and services — exists in the system.

Reallocating it, of course, would be a major shock, since people with lots of money right now really don’t want to let go of it. This is amplified by people’s tendency to think of our rules of who “rightfully owns” what to be laws of nature: that to move money away from large stores of wealth means people taking from people what’s rightfully theirs.


It’s important to recognize how mutable our ideas of property have been over time. People can “own” things in many senses: You own your car in a different sense than you own your body, or your thoughts, or your time, or your family. Your kidneys cannot (for the moment) be seized from you by a court to pay a debt, but your car can.

Historically, different kinds of ownership came with different rights. For example, most British homeownership doesn’t include land ownership, and land ownership in turn doesn’t include a right to bar passage over it: The public right to walk over common paths is considered to take precedence. Likewise, American land ownership doesn’t automatically include either mineral rights or control over the airspace. Most traditional land-ownership systems don’t actually have land owned by individuals at all; land is typically owned by clans, families, or communities, and is considered inalienable property in much the same way that body parts are.

Even the kinds of things which can be owned have changed. “Intellectual property,” for example, is an 18th-century innovation, and its boundaries have shifted even in our lifetime. Slaves, on the other hand, are no longer considered a legitimate kind of property by most societies.

Within most Western societies today, anything which can be traded on a market can be owned as personal property, and is considered to convey roughly the same kind of ownership. In fact, much of the argument against allowing payment for organ donors is that if this were possible, organs would de facto become market goods and could be seized in payment of a debt, either directly or by putting people in a position where selling their organs was their only option!

What all of this means is that at the moment, even very large concentrations of wealth can be considered “personal property” in the same sense that a toothbrush is. We could imagine ways to change this — for example, what if we effectively separated goods into three tiers: things necessary for survival (the things whose lack would push you into the far left of the graph), things you could own beyond that, and chunks of wealth too large for any single person to meaningfully “own”? Someone ending up with an accumulation of wealth above this threshold, either individually or via a company, might find themselves owning it in the sense of being the custodian of a public trust, or alternatively they might not own it in any meaningful sense.

I can tell you from my own experience in Silicon Valley, having lived through the booms and busts, that one “owns” large stockholdings in a very different sense than one “owns” one’s home. It’s just this thing that you have, and it’s important for reasons like being able to survive a bigger financial shock (a cancer diagnosis won’t automatically ruin you!), or being able to guarantee your children have access to the elite education systems which grant them access to tier 1 and 2 jobs, or (ultimately) access to social and political power.

Critique #3: People Will Lose Their Dignity

The third critique of a universal basic income is one whose importance, I think, has been grossly underestimated: that people would lose their dignity by taking a basic income instead of working, that they would become parasites on society instead of contributing members.

I first heard this critique from Rob Gordon, and thinking about it deeply was one of the main things that caused me to move to my current job. To quote one of my notes about it from an earlier article, one of the key things we get from jobs is dignity:

“a validated sense that one’s contributions are valued by one’s community. One’s community, including anything from one’s family, to one’s neighbors, to one’s colleagues; valued, in that people see them as important; one’s contributions, as distinct from one’s intrinsic self, because if you are putting a lot of effort into something it’s important to people to feel that this is recognized; and validated, because people can recognize bullshit, and real valuation feels like it.”

One of the major operational limits of social welfare programs like Medicare and food stamps is that many people eligible for them refuse to accept them. They view participating in such programs as not just an admission of failure, but an act of taking from society while not giving to it. People want to contribute to their society, and they want that contribution to be seen and valued. Note also that there are two important kinds of social valuation people want: a basic level of being valued as a person and a second level being valued for effort and hard work.


It’s worth taking a moment now to look at job guarantees, since they are an attempt to answer this third critique, among others, and highlight things that do and don’t work.

Rather than giving everyone a dividend like a universal basic income, a job guarantee provides access to a job which will pay a livable amount. It uses the same sources of funding that would power a basic income.

The job guarantee is a way to directly address the first and third critiques of the basic income. It ensures that everyone has some kind of work, which could provide them with dignity. Also, it ensures that the “shit jobs” continue to be filled. (The second critique on funding sources applies in exactly the same way.)

But on further examination, job guarantees seem to be an inadequate solution.

Job guarantees answer the first critique that people won’t want to work by deliberately leaving the underlying problem intact. Anyone who has to rely on a job guarantee has no better job available; therefore, all the “shit jobs” can be made into job guarantee jobs, which continue to be low-paying.

So a job guarantee “solves” the problem by maintaining precisely the imbalances which exist in our society today. People with access to what are currently type #1, #2, or #3 jobs are largely unaffected by the change, and type #4 “terrible” jobs are now “guaranteed” jobs.

We should be immediately suspicious of this for several reasons. First of all, if the jobs that are already type #4 were enough to give everyone who wanted a job a job, why isn’t everyone already working in them? Really, a job guarantee would amount to two changes in this situation:

  • A mandated livable wage for type #4 jobs.
  • A potential increase in the pool of type #4 jobs.

We’ve seen an attempt at this kind of solution before; the Soviet Union loved it. I remember my great-aunt telling me of the time that she was invited to visit a Soviet bottle factory as a visiting foreigner. She could see how raw materials were melted, poured into molds, cooled, stacked, smashed with hammers, and carted right back to the beginning of the assembly line. You see, bottle production levels had been mandated by the government, but nobody needed these bottles. They couldn’t simply shut down the factory (because people would lose their jobs), and they couldn’t repurpose it (because the idea of being responsive to actual demand for products was somehow beyond the Politburo’s ability to conceive), so people were deliberately engaging in pointless work.

In the modern era, one would hope that the inability to repurpose a factory wouldn’t be a limiting factor, but we’re already far into a state where the number of total manufacturing jobs needed is far lower than the number of people who want to work in them. It’s not that we can’t think of repurposing a factory — it’s that there’s nothing obvious to repurpose it to do.

And this leads right into how job guarantees could easily fail to address the third critique on dignity as well. If you create a job which is obviously pointless make-work, it doesn’t give anyone dignity. It just highlights for them how pointless and demeaning what they’re doing is.

Basically, a job guarantee is just a living minimum wage coupled with a lot of gyrations to put people in pointless jobs. It could be an easier sell than a universal basic income, since it maintains current power hierarchies, but by doing so it eliminates a lot of its own purpose.


The above indicates, I think, that both a basic income and job guarantees are inadequate solutions. That doesn’t mean they’re bad ideas — it just means that we’re still grasping toward a real solution.

I have a few preliminary ideas about what a better solution might look like. To explain, I want to go back to the question of the dignity we get from our work.

We get dignity from our work because what we do is important to our community, is seen as a valued contribution to it. But we’re prone to confusing our work with our job. The fact is that a lot of us contribute actively to our communities in ways we don’t get paid for: We raise families, we volunteer, we help people resolve their differences, we organize events, we listen to each other and provide emotional support, we pay attention to each other’s needs.

These contributions are crucially important to our survival as a society. They’re things we’ve been doing forever; it’s only in the past 150 years or so, as more and more people have started spending all of their time doing “traditional” wage-labor jobs, that the focus has shifted away from these other types of work.

The monetization of labor that came with the industrial revolution led to an odd twist: We came to only value labor in terms of how much it paid. Suddenly, raising children or making food became not just devalued, but not even thought of as labor. (It was only in the 1960s that economists started talking about “household labor” again at all!) The work of project-managing a household — of bringing people together and monitoring their needs, dealing with unhappy and unpleasant people — was dropped from our definition of “labor” until the 1980s, when the idea of “emotional labor” very slowly started to work its way into people’s consciousness . And it’s still a little-understood idea!

But if you have ever had to do any of this, you’re very aware of how much work it is, and of how important it is. Or, if you’ve ever suddenly found yourself in a situation where the people who were doing it stopped, you probably found out pretty quickly how valuable it was.

These forms of “invisible labor” are afforded neither the compensation nor the dignity that we give to “visible” kinds of labor paid an hourly wage or a salary.

The automation changes that we started this discussion with are eliminating the need for a lot of the jobs that we associate with visible labor, but not the need for invisible jobs. We all find ourselves doing more and more emotional labor, household labor, probably several other kinds of labor that I don’t yet have the words to name, but we are neither paid for them nor do we receive social recognition or dignity for doing them.


Maybe this is what we need to change.

Consider: If the production of material goods is increasingly handled by robots, then people can spend more of their time on these invisible labor tasks. Unlike the traditional “substitution” argument for jobs — “oh, those laid-off people will just retrain” — which we now hopefully all see as transparent bullshit, these are the tasks which people are doing when they aren’t spending time on their jobs.

So step one: Before we try either a basic income or a job guarantee or anything else, let’s put a serious effort into giving names and recognition to these other kinds of work. Let’s make our invisible labor visible.

Now let’s return to the three purposes of a job, this time considering both the traditional and invisible kinds:

  1. To make things we need — both traditional goods and services, and invisible goods and services like food, social networks, children, and so on.
  2. To allocate resources (via things like wages).
  3. To achieve individual meaning in our lives.

It’s suddenly clear that, if we make the invisible visible, there are more than enough jobs to be had which satisfy objectives #1 and #3. The only catch is that a lot of these invisible jobs, which are now suddenly the majority of work that people are doing, don’t satisfy objective #2. Compare that to where we started — where jobs were vanishing, people had no work whatsoever, and neither #2 nor #3 was satisfied — and we suddenly have a different problem to solve!

If we do make our invisible labor visible, then the problem of dignity is addressed directly, as is the problem of a lack of jobs. Our problem is instead to ensure that the previously invisible labor is financially compensated.

This is a very different problem, with many different possible solutions. We could try for a wholly market-based solution, with people paying each other for their emotional labor and the like. But my gut feeling is that this would be unsatisfactory, both in terms of how well it would work (if your whole community is broke but people are doing lots of this labor for each other, you can’t easily pay them in money) and in terms of how people would feel about it. We could also go to a more basic income-shaped solution, where people are getting paid some fixed amount because of the labor they’re undoubtedly doing.

Or (as I suspect is wisest) we could start thinking about hybrid solutions. For example, what if we had universal access to minimum survival resources (to avoid people falling off the left edge of the cliff, and all the net social losses that come with that), but also a job guarantee tied to this previously invisible labor? Or alternatively, what if we just had a job guarantee that if you do this job, you will be paid for it? Or what if we engaged in some other wealth distribution mechanism to ensure that communities had enough resources to pay people (in money or otherwise) for their emotional labor, and then left the rest to the market?

There are a lot of possibilities here. I have a strong suspicion that if we can solve the problem of making invisible labor visible, we will change the context of our broader problem in a way that solutions free of the flaws of either a universal basic income or a job guarantee will become possible.

And then, maybe, we can enjoy the fruits of a society not on the verge of material starvation.


Postscript: Further Thoughts

Almost immediately after this article was first posted, it met with a flood of very thoughtful responses. A few points seem worth highlighting.

The dynamics of a small basic income versus a large basic income are very different. Specifically, a basic income large enough that someone no longer has to work at all has a radically different economic effect than one that keeps someone working at even a low-paying job off the left edge of the cliff. Honestly, the small one is probably better for all concerned, since the large one ends up with feedback loops so long as there are lots of jobs that have to be done by humans: wages start rising, but basic income costs stay fixed, and suddenly all the resources in a society are going into those things, with inflation risks as well.

I was running some simulations in my head last night of how you might deploy a basic income. The first conclusion was that suddenly deploying a large one would be catastrophic: A lot of people would say “see ya” to jobs they hate, but the ability of employers to raise wages enough to retain them wouldn’t shift nearly as fast. Suddenly a lot of critical jobs would be unstaffed. That doesn’t have to happen for very long at all (by which I mean, 48 to 72 hours) to cause things like a collapse of the food logistics network, which is Very Bad. But this shouldn’t be too surprising; a basic rule of plumbing is “never open a valve quickly, especially if it’s under a lot of pressure,” and it applies pretty well to economies as well.

On the other hand, a small basic income is much closer to what’s already been field-tested in places like Manitoba, and could easily solve a lot of the “falling off the left edge” problems without any of these risks. So starting the knob somewhere around there (imagine something in the ballpark of $500 to $1000 per month) seems much wiser.

A point brought up by Jennifer Freeman and Anne Faye is that the third critique of a basic income related to dignity may actually be entirely false for a different reason: Existing social support systems, especially in the U.S., have been deliberately engineered to rob people of dignity. But experiments in systems not engineered to do so (including the few prior basic income tests, and various paid-leave systems tested in Europe) suggest that the “loss of dignity” problem may not manifest at all. People start doing precisely the kinds of invisible labor described here, and derive worth and value from it, with no intervention required. This is a promising signal! (It also highlights the extent to which American social welfare systems have humiliation embedded in them as a deliberate design parameter.)

The critiques of job guarantees I listed may have some interesting counterpoints as well. Robert Kennedy pointed out how the U.S. jobs programs during the Great Depression didn’t suffer from these issues at all, but that this may have had to do with their short-term nature. It’s definitely true that investment in infrastructure (everything from roads to art — things that create lasting value) is almost always a good bet for a society, so if job guarantees take that form they could work out quite well.

On the other hand, the critique I didn’t spend as much time on has to do with a job guarantee excluding people who can’t work. This is very serious, and probably should have gotten more text. One thing to compare it to is President Bill Clinton’s 1994 “welfare reform,” which had catastrophic consequences for the people previously using these services who suffered from illnesses, family obligations, or other things which prevented them from working. The biggest problem with any system which rewards work in such a fashion is that it tends to also penalize an inability to work, and only people “sufficiently documented” as unable to work are at all excluded. (Such documentation is very hard to get, and having it often prompts other negative consequences, such as complications with health insurance.) Honestly, this is the top reason why I think job guarantees are a very, very bad idea.


Lots of people have also asked why I didn’t describe a basic income as the obvious basis for any solution. I’m not fully convinced that it’s obvious yet, but I do think at this point that something roughly basic income-shaped is where the best solution is going to be. The optimal configuration is probably going to be a lot more complex than a straight basic income, though. For example, there’s reason to provide some things in kind rather than as cash (medicine, for example, where inflationary pressures could be especially high and economies of scale are substantial). There are a lot of good questions about what the best level of a basic income would be, and how that would vary regionally. There’s the notion of infrastructure-producing jobs (using “infrastructure” in the broadest possible sense) as a partial alternative, which has a lot of long-term value if it can be done right. And there are all sorts of complicated problems tied in with political borders, from “who has access to this” to “how do you prevent capital flight.” There’s the question of systematically addressing the dignity issue and whether that’s required, and of just what a deployment would look like — a hugely complicated question that I didn’t even touch on here. And all of that, of course, is predicated on building a social consensus around doing it, which is a nontrivial problem given several of the issues noted above.

In short: I continue to think this is very worth exploring and testing, but I’m not at the point where I’m ready to say, “Oh yes, obviously this is the right idea.”