manufacturing

Why “bullshit jobs” exist

PNRJ inequality, macroeconomics, market failure, mathematical models, public policy , , , , , ,

May 24 JDN 2461185

Content warning: Algebra

In 2018, David Graeber published a book called Bullshit Jobs, positing that the transition of our economy from industrial manufacturing to ‘post-industrial’ services was in fact largely a transition from meaningful jobs that do actual work to meaningless jobs that employ people without contributing to society.

He made the point vividly in this article for Strike magazine:

For instance: in our society, there seems a general rule that, the more obviously one’s work benefits other people, the less one is likely to be paid for it. Again, an objective measure is hard to find, but one easy way to get a sense is to ask: what would happen were this entire class of people to simply disappear? Say what you like about nurses, garbage collectors, or mechanics, it’s obvious that were they to vanish in a puff of smoke, the results would be immediate and catastrophic. A world without teachers or dock-workers would soon be in trouble, and even one without science fiction writers or ska musicians would clearly be a lesser place. It’s not entirely clear how humanity would suffer were all private equity CEOs, lobbyists, PR researchers, actuaries, telemarketers, bailiffs or legal consultants to similarly vanish. (Many suspect it might markedly improve.) Yet apart from a handful of well-touted exceptions (doctors), the rule holds surprisingly well.

I think his claim is a little overstated, but the basic pattern does seem valid: Less and less labor is spent actually making things (or even really doing things) and more and more is spent on selling things and advertising things and suing over things.

While the pattern seems genuine, it’s a little hard to verify empirically, as the usual BLS categories include both useful and useless work in the same categories: “Sales and office occupations” includes far too many things to be useful.

But I think Graeber’s explanations of where the pattern comes from are pretty much totally wrong. He attributes it to “office feudalism” where bosses want to act like lords and “make-work” from government policy trying to create jobs without regard for whether those jobs are useful. I’m not saying these things never happen, but it’s not enough to explain a massive economic transition.

Rather, with a little bit of economic theory, I think it’s not hard to see that these “bullshit jobs” are actually rent-seeking; they don’t produce anything for society, but they are genuinely profitable for the companies that hire them—and thus it’s perfectly rational for a profit-maximizing corporation to behave in this way.

Here’s a simple model to demonstrate the idea.

Suppose there are two corporations, Zero Inc. and One Corp, which both produce cars.

Each factory worker produces r cars. (Presumably they really produce some fraction 1/n of n r cars, but it works out the same.) The labor productivity of factory workers is thus r.

Each factory worker receives a wage wf, which is outside the control of these two corporations (it’s set by a larger labor market).

This means that the marginal cost of producing a car is wf/r. As productivity increases, producing cars becomes cheaper.

The two corporations are in Cournot competition, so they face a total demand for cars that looks like this:

p = a – b (q0 + q1)

where q0 is how many cars Zero produces and q1 is how many cars One produces, and a and b are parameters. This is exogenous; it’s determined by broader market conditions, not by anything the corporations can control.

I’ll spare you the algebra, but this has a well-known equilibrium solution; the two corporations produce the same amount and charge the same price:

q0 = q1 = (a – wf/r)/(3b)
p = (a + 2wf/r)/3

Their profits are also the same, naturally:

F0 = F1 = (p – wf/r)q0 = (a – wf/r)2/(9b)

The number of workers employed making cars is:

L = q/r = 2(a – wf/r)/(3br)

Because the numerator increases with r but so does the denominator, it isn’t obvious whether this is increasing or decreasing in r; more productivity may result in either more or fewer factory workers employed.

By taking the derivative, this can be shown:

dL/dr > 0 iff a > 2w/r

This means that at relatively low levels of productivity, more productivity will increase employment of factory workers; but at high levels of productivity, it will decrease it.

Since profits are proportional to (a – wf/r)2, the higher profits are, the more likely it is that additional productivity will reduce employment of factory workers.

But what if factory workers aren’t the only workers each corporation can employ?

Suppose now that there is another kind of worker they can employ, salesmen.

Zero employs s0 salesmen, while One employs s1. Salesmen are paid a wage ws.

Salesmen don’t actually make anything. They produce no real output. Instead, they allow each firm to take a larger share of the market.

To model this, I’ll use a contest function (I used a similar model in some posts in 2020), where the actual quantity each corporation produces and sells is proportional to the relative number of salesmen employed at each corporation:

q0 = (s0)/(s0 + s1) q

I’ll assume that the total amount of cars produced and sold is the same as before. (Dropping this assumption makes the model much more complicated and hard to solve, without significantly changing the overall conclusions.)

Each corporation’s profits now depend on both how many cars they produce and how many salesmen they hire:

F0 = p q0 – wf/r q0 – ws s0

Again I will spare you the algebra, but it comes out like this:

s0 = s1 = (a – wf/r)2 / (18 b ws)

Unlike employment of factory workers, which may increase or decrease as productivity increases, employment of salesmen always increases.

Thus, as productivity gets higher and higher, we should expect employment of factory workers (“real jobs”) decreasing as employment of salesmen (“bullshit jobs”) increases—even though corporations are being completely rational and maximizing their profits.

Let’s put some numbers on this to make the example more concrete.

Suppose at the start that wf = 1, ws = 2, a = 10, b = 1, and r = 1. (Consider production in cars per day, wages as annual salaries, and money in units of $10,000.)

Then each corporation produces 3 cars per day:

q0 = q1 = (10 – 1/1)/(3*1) = 3

The price of a car is $40,000:

p = (10 + 2*1/1)/3 = 4

The number of factory workers employed is 6:

L = q/r = (3+3)/1

The number of salesmen employed is 4.5 (maybe one is employed half-time):

s = s0 + s1 = (10 – 1/1)2 / (9*1*2) = 81/18 = 4.5

So the majority of workers are factory workers.

But now suppose that productivity greatly increases, to r=10, with everything else the same:

The number of cars produced at each corporation per day increases to 3.3:

q0 = q1 = (10 – 1/10)/(3*1) = 3.3

The price of a car falls to $34,000:

p = (10 + 2*1/10)/3 = 3.4

The number of factory workers employed falls dramatically; there isn’t even one full-time job available, only part-time jobs:

L = (3.3+3.3)/10 = 0.66

But the number of salesmen increases to 5.445 (one additional full-time worker):

s = (10 – 1/10)2 / (9*1*2) = 5.445

Thus, while factory workers used to be 6/(6+4.5) = 57% of the workforce, now they are only 0.66/(0.66+5.445) = 10.8% of the workforce. And this happened because they got more productive!

This is of course a very simple model, but the basic pattern fits:

The share of employment that’s in manufacturing has greatly declined since 2000:

Yet manufacturing output hasn’t really changed:

The reason is that labor productivity has continued to rise:

In my model I assumed that factory worker wages haven’t changed, and, well, they really haven’t, despite large increases in productivity. In real terms, a factory worker today makes about 70% more than one in 1950, despite producing nearly six times as much output.

Producing the same amount of stuff now requires fewer people—so fewer people are employed in making stuff. They then end up employed somewhere else; and one of the more profitable ways to employ them these days is in rent-seeking activities like sales and advertising.

What should we do about this?

I already said what I wanted to do years ago: Tax advertising.

Conflict without shared reality

PNRJ core principles, current events, ethics, inequality, politics, tribal paradigm , , , , , , , , , , , , , , , , , , , , , , , ,

Aug 17 JDN 2460905

Donald Trump has federalized the police in Washington D.C. and deployed the National Guard. He claims he is doing this in response to a public safety emergency and crime that is “out of control”.

Crime rates in Washington, D.C. are declining and overall at their lowest level in 30 years. Its violent crime rate has not been this low since the 1960s.

By any objective standard, there is no emergency here. Crime in D.C. is not by any means out of control.

Indeed, across the United States, homicide rates are as low as they have been in 60 years.

But we do not live in a world where politics is based on objective truth.

We live in a world where the public perception of reality itself is shaped by the political narrative.

One of the first things that authoritarians do to control these narratives is try to make their followers distrust objective sources. I watch in disgust as not simply the Babylon Bee (which is a right-wing satire site that tries really hard to be funny but never quite manages it) but even the Atlantic (a mainstream news outlet generally considered credible) feeds—in multiple articles—into this dangerous lie that crime is increasing and the official statistics are somehow misleading us about that.

Of course the Atlantic‘s take is much more nuanced; but quite frankly, now is not the time for nuance. A fascist is trying to take over our government, and he needs to be resisted at every turn by every means possible. You need to be calling him out on every single lie he makes—yes, every single one, I know there are a lot of them, and that’s kind of the point—rather than trying to find alternative framings on which maybe part of what he said could somehow be construed as reasonable from a certain point of view. Every time you make Trump sound more reasonable than he is—and mainstream news outlets have done this literally hundreds of times—you are pushing America closer to fascism.

I really don’t know what to do here.

It is impossible to resolve conflicts when they are not based on shared reality.

No policy can solve a crime wave that doesn’t exist. No trade agreement can stop unfair trading practices that aren’t happening. Nothing can stop vaccines from causing autism that they already don’t cause. There is no way to fix problems when those problems are completely imaginary.

I used to think that political conflict was about different values which had to be balanced against one another: Liberty versus security, efficiency versus equality, justice versus mercy. I thought that we all agreed on the basic facts and even most of the values, and were just disagreeing about how to weigh certain values over others.

Maybe I was simply naive; maybe it’s never been like that. But it certainly isn’t right now. We aren’t disagreeing about what should be done; we are disagreeing about what is happening in front of our eyes. We don’t simply have different priorities or even different values; it’s like we are living in different worlds.

I have read, e.g. by Jonathan Haidt, that conservatives largely understand what liberals want, but liberals don’t really understand what conservatives want. (I would like to take one of the tests they use in these experiments, see how I actually do; but I’ve never been able to find one.)

Haidt’s particular argument seems to be that liberals don’t “understand” the “moral dimensions” of loyalty, authority, and sanctity, because we only “understand” harm and fairness as the basis of morality. But just because someone says something is morally relevant, that doesn’t mean it is morally relevant! And indeed, based on more or less the entirety of ethical philosophy, I can say that harm and fairness are morality, and the others simply aren’t. They are distortions of morality, they are inherently evil, and we are right to oppose them at every turn. Loyalty, authority, and sanctity are what fed Nazi Germany and the Spanish Inquisition.

This claim that liberals don’t understand conservatives has always seemed very odd to me: I feel like I have a pretty clear idea what conservatives want, it’s just that what they want is terrible: Kick out the immigrants, take money from the poor and give it to the rich, and put rich straight Christian White men back in charge of everything. (I mean, really, if that’s not what they want, why do they keep voting for people who do it? Revealed preferences, people!)

Or, more sympathetically: They want to go back to a nostalgia-tinted vision of the 1950s and 1960s in which it felt like things were going well for our country—because they were blissfully ignorant of all the violence and injustice in the world. No, thank you, Black people and queer people do not want to go back to how we were treated in the 1950s—when segregation was legal and Alan Turing was chemically castrated. (And they also don’t seem to grasp that among the things that did make some things go relatively well in that period were unions, antitrust law and progressive taxes, which conservatives now fight against at every turn.)

But I think maybe part of what’s actually happening here is that a lot of conservatives actually “want” things that literally don’t make sense, because they rest upon assumptions about the world that simply aren’t true.

They want to end “out of control” crime that is the lowest it’s been in decades.

They want to stop schools from teaching things that they already aren’t teaching.

They want the immigrants to stop bringing drugs and crime that they aren’t bringing.

They want LGBT people to stop converting their children, which we already don’t and couldn’t. (And then they want to do their own conversions in the other direction—which also don’t work, but cause tremendous harm.)

They want liberal professors to stop indoctrinating their students in ways we already aren’t and can’t. (If we could indoctrinate our students, don’t you think we’d at least make them read the syllabus?)

They want to cut government spending by eliminating “waste” and “fraud” that are trivial amounts, without cutting the things that are actually expensive, like Social Security, Medicare, and the military. They think we can balance the budget without cutting these things or raising taxes—which is just literally mathematically impossible.

They want to close off trade to bring back jobs that were sent offshore—but those jobs weren’t sent offshore, they were replaced by robots. (US manufacturing output is near its highest ever, even though manufacturing employment is half what it once was.)


And meanwhile, there’s a bunch of real problems that aren’t getting addressed: Soaring inequality, a dysfunctional healthcare system, climate change, the economic upheaval of AI—and they either don’t care about those, aren’t paying attention to them, or don’t even believe they exist.

It feels a bit like this:

You walk into a room and someone points a gun at you, shouting “Drop the weapon!” but you’re not carrying a weapon. And you show your hands, and try to explain that you don’t have a weapon, but they just keep shouting “Drop the weapon!” over and over again. Someone else has already convinced them that you have a weapon, and they expect you to drop that weapon, and nothing you say can change their mind about this.

What exactly should you do in that situation?

How do you avoid getting shot?

Do you drop something else and say it’s the weapon (make some kind of minor concession that looks vaguely like what they asked for)? Do you try to convince them that you have a right to the weapon (accept their false premise but try to negotiate around it)? Do you just run away (leave the country?)? Do you double down and try even harder to convince them that you really, truly, have no weapon?

I’m not saying that everyone on the left has a completely accurate picture of reality; there are clearly a lot of misconceptions on this side of the aisle as well. But at least among the mainstream center left, there seems to be a respect for objective statistics and a generally accurate perception of how the world works—the “reality-based community”. Sometimes liberals make mistakes, have bad ideas, or even tell lies; but I don’t hear a lot of liberals trying to fix problems that don’t exist or asking for the government budget to be changed in ways that violate basic arithmetic.

I really don’t know what do here, though.

How do you change people’s minds when they won’t even agree on the basic facts?