David Graeber introduced the idea of "Bullshit Jobs" as a primary focus of inquiry to public discourse in On the Phenomenon of Bullshit Jobs: A Work Rant, and the book Bullshit Jobs, but unfortunately with a tendency to make strong categorical statements even when his categories were not very good.¹ Even more unfortunately, he's dead now, so I can't write to him and persuade him to fix this problem in his own name.

In The Debtors' Revolt I explained some of the structural reasons to expect such jobs to exist, but thought it might be a good idea to simply enumerate some of the categories, after which it should be easier to see that quite a lot of jobs are in aggregate intentionally economically wasted effort, even when they locally benefit some counterparty.

Examples follow.

Your job might be to threaten trouble for someone else, or to correspondingly protect someone from that potential trouble. Much of what soldiers, lawyers, politicians, lobbyists, and accountants specializing in taxation or fraud do is in this category. If you are religious and take your religion literally, you should probably agree that the clergy of most other religions - i.e. most clergy - are doing this. If you do not take it literally then you should probably agree that the clergy of your own religion are also doing this.

Your job might be to persuade people to buy Coca-Cola, or some analogous product that is useless or harmful. Part of your work is zero-sum, persuading people to purchase your product rather than a nearly identical alternative offered at a lower price. Part of your work is directly harmful, persuading people to buy a useless or harmful product, instead of spending money on something more useful to them, or enjoying more leisure.

Your job might be to help produce or distribute such a product.

Your job might be to occupy preexisting levels of intermediation in a large stable organization, use your position of trust to create new levels of intermediation, create a plausible story that those new levels of intermediation are responsible for the output of the people doing the actual work, structure information flows so that it is hard to disintermediate you, and blackball any potential institutional partners without a similar class of intermediaries you can relate to.² You might work for a separate organization that specializes in such intermediation. For some details, see Moral Mazes, and Parkinson’s Law and the Ideology of Statistics.

Your job might be to produce useless information products that the prior class can pretend to intermediate. For example, some managers may employ people producing reports for the sole purpose of allowing the manager to be able to feel - and seem to other managers like they feel - that the things they are saying at meetings are informed by those reports, and therefore that those reports are part of the process by which whatever gets done at their company gets done. Likewise, universities and grantmaking organizations justify their existence in part on the basis of published research by the academics they fund, who in in many contexts have to pay subordinate researchers as part of the performance, whether or not any meaningful knowledge is produced. For examples, see the Replication Crisis, Sokal Hoax, and Grievance Studies Affair.

Your job might be in whole or part some other sort of theater that serves entirely to increase the perceived importance of whoever hired you, like a liveried footman or a receptionist.

Your job might be to gatekeep access to a good or service, again in a manner that is designed not to allocate resources efficiently with minimal overhead, but to create jobs for gatekeepers.

Your job might be to perform a ritualized imitation of a service that would be useful if genuine. Bryan Caplan's book The Case Against Education is a good case study; formal secondary and post-secondary education appears to mostly be like this, at least in the US. Much medicine as performed by licensed physicans and similarly licensed professionals in the US likewise appears to provide no measuranle net health benefit; Robin Hanson has documented much of the case for this; Cut Medicine in Half and Medical Doubts OpEd cover much of the relevant academic evidence. Such imitations waste the time not only of the providers but of the supposed beneficiaries.

Your job might be to perform a perfectly necessary part of the process by which a perfectly legitimate good or service is produced, but under explicit or tacit work rules that increase labor's share of productivity, not by increasing compensation per hour worked, but by wastefully multiplying the number of hours worked necessary to produce the product.

Other jobs fall somewhere in the overlap between some of these categories. For instance, "content creators" frequently produce an apparent information good optimized to capture attention rather than to inform, in order to sell some of the attention they capture to people trying to persuade the viewers to purchase some other product.

Even if your job is a perfectly necessary part of the process by which a perfectly legitimate good or service is produced or distributed, under reasonably efficient work rules, some share of your output may be used for systematically wasteful purposes, and is therefore part of the total waste caused by bullshit jobs. For instance, a friend of mine who works at an Amazon warehouse probably spends a significant amount of his time helping to efficiently distribute useless or harmful products, and to distribute otherwise useful products to people engaged in more directly wasteful activities, e.g. a computer to be used by a manager to create a PowerPoint presentation justifying the hiring of a junior manager to take credit for some of the things the manager's existing direct reports do.

Likewise some people produce things that are subsidized directly or indirectly to create jobs, so that those things are overproduced and have to be stockpiled (e.g. the national cheese caves) or shoehorned into products that no one would make that way if not for the subsidy regime, and are somewhat worse because of it (e.g. high fructose corn syrup in foods, and ethanol in gasoline).

---

1. In a comment on B.P.S.'s Substack article I Used to Believe in STEM, I wrote:
   
   I think the underlying problem is that a tremendous share of economically validated activity in the postwar American system - especially when weighted by money moved per unit of time demanded - is bullshit. This is best understood in the context of an ongoing transition from the behavioral complex of meeting exogenous performance demands (e.g. food scarcity, hot war), to the behavioral complex of patronage coalitions seeking rents. [...]
   
   The correct frame of analysis is unfortunately antimemetic [...], in part because people are directly discouraged from thinking about it, but probably mostly because these coalitions have a strong survival incentive to confound any attempt to cleanly categorize them.
   
   One way these coalitions confuse people is to coopt any attempt to represent and advocate for a class of people doing real work by including both real and simulacral versions of it in their coalition.
   
   A straightforward example is the "labor" movement, which may have started as an attempt by productive workers to organize collective bargaining power to mitigate unjust exploitation by state-backed labor monopsonists , but gradually became just another sort of political pressure group, and for decades it's a commonplace idea that union rules are frequently a sort of goldbricking that require extraneous, i.e. bullshit, "labor."
   
   Likewise, the term "scientist" derives its legitimacy from things like the activities of the Royal Society, but now mostly describes paper-pushing careerist academics whose published output interests no one, and exists mostly to satisfy institutional demands for publications as a simulacral metric of productivity.
   
   Nominally for-profit businesses include low-margin businesses in competitive fields, entrepreneurial innovators developing new products that displace old monopolies. Both categories can only employ people because they are, or are expected to be, profitable. But the category also includes implicitly state-backed corporations that can collectively count on bailouts to keep them long-run financially sustainable. They're profitable because they employ so many people.
   
   Another way to confuse people is to coopt attempts to discuss the problem by promoting simplistic distinctions that fail to capture the true problem. Grifters in STEM benefit from the promotion of the STEM vs Humanities dichotomy as a substitute for the general distinction between needful and pretend work, and grifters in the Humanities can likewise recruit well-intentioned allies on the basis of the mistakes built into the pro-STEM point of view.
   
   David Graeber's otherwise excellent essay and book "Bullshit Jobs" suffer from these sorts of confusion. He can muster plenty of persuasive evidence that bullshit work exists and is atrociously common, but is constantly tempted to try to cleanly categorize particular *sorts of jobs* as bullshit and each time gets it wrong, because the bullshit strategic behavioral complex is actively optimizing to make itself hard to cleanly distinguish from the outside with simple distinctions. ↩︎
2. From the comments:
   
   Elon Musk's acquisition of Twitter provided interesting data [...] He laid off most of the staff. Ad revenues seem to have declined. My understanding is that usage hasn't declined much, nor is it plausible that big institutional advertisers forgot that Twitter exists. But this is well explained by the mechanism of "blackball any potential institutional partners without a similar class of intermediaries you can relate to."
   
   Basically, big advertisers stopped placing ads on Twitter because it was no longer a jobs program for the kind of people who decide where big advertisers advertise. Likewise, I've heard informally that companies that aren't trying to become jobs program for the kinds of people who work in enterprise sales, seem to have difficulty closing enterprise sales, even when they make the only product in the category, and the line employees are begging management for the product, and the CEO of the customer wants to close the deal. ↩︎