Bullshit Jobs.

The essay that became a book that became a number

August 2013. David Graeber, anthropologist, author of Debt: The First 5,000 Years, future intellectual cornerstone of Occupy Wall Street, publishes a short essay in Strike! magazine. It is titled On the Phenomenon of Bullshit Jobs. The piece argues that capitalism, which was supposed to abolish drudgery through automation, has instead invented vast quantities of new drudgery — work that the worker themselves privately believes should not exist — and that this is a structural feature of the modern economy, not an aberration.

The essay went viral the way important essays go viral — quietly, persistently, in inboxes and break rooms and the comment threads of HR forums. Within months, YouGov ran the obvious survey. 37% of UK workers said their job did not make a "meaningful contribution to the world." Another 13% were unsure. The Netherlands replicated the finding shortly afterwards with similar numbers. The phenomenon was not Graeber's imagination.

The book — Bullshit Jobs: A Theory — followed in 2018, drawing on hundreds of testimonies Graeber had collected from workers in the intervening years. The methodology is anthropological, not econometric. The diagnostic is the worker's own private judgement, made under conditions of professional pretence. The argument is that this private judgement is what an honest economy would ledger.

The five categories

Graeber's taxonomy of bullshit work, drawn from the testimonies. Each category names a structural reason the job exists despite the worker's correct sense that it should not.

1. Flunkies. Jobs that exist to make someone else look or feel important. The receptionist at a firm that does not need a receptionist. The personal assistant whose actual function is to confer status by their existence. The junior team retained at headcount to fill a floor of the building. The work performed is incidental; the structural function is hierarchy display.
2. Goons. Jobs whose only justification is that other people employ them — corporate lawyers, lobbyists, certain categories of PR, telemarketing, high-frequency-trading desks. Each side employs goons because the other side does. If both sides disarmed, the work would not need to exist. The category is structurally zero-sum — useful for the employer in the local game, useless to the economy as a whole.
3. Duct-tapers. Jobs that exist to patch over a structural problem that should not exist. The engineer whose actual job is to work around a manager's bad decisions. The translator employed because the firm refuses to standardise its communication. The customer service representative whose principal function is to absorb anger generated by a deliberately confusing pricing model. Useful in the moment; embarrassing on inspection.
4. Box-tickers. Jobs that exist so an organisation can claim it is doing something it is not. Compliance theatre. Diversity reports written to be filed, not read. Strategy decks produced to demonstrate that strategy is happening. The form is the substance; the substance is the appearance. Bureaucratic mass produced for the sake of producing it.
5. Taskmasters. Jobs that exist to manage people who do not need managing, or to invent bullshit work for others. The mid-tier manager whose subordinates know what they are doing. The corporate trainer whose programmes actively impede the work. The layer of supervision that generates more friction than it removes. The taskmaster is both the cause and the consequence of the other categories; it is the metabolic engine of managerial feudalism.

The structural function of a bullshit job
is not what it produces — it is what it signals.

Why this is not a market mistake

The natural objection is that markets should eliminate bullshit jobs ruthlessly — any firm carrying useless headcount should be out-competed by one that does not. Graeber takes the objection seriously and gives the most important answer in the book: the modern firm, especially in its financialised and managerial form, is not principally a profit-maximising machine. It is a hierarchy that also has revenue.

The argument is anthropological. In feudal economies, economic power and political status are continuous; a lord accrues retainers because retainers are the visible form of being a lord. The retainers do not need to be productive; their existence is the productive output of having retainers. Graeber argues that large parts of the contemporary corporate economy now operate on this logic — managers accrue underlings because underlings are how managerial importance is performed to other managers, the board, the analyst calls, the conference circuit. The underlings, in turn, accrue their own underlings. The pyramid is the product.

The financial sector is the clearest case in Graeber's treatment. A trading desk is not a productivity-maximising entity in the textbook sense; it is a status hierarchy with a P&L attached. The number of analysts a managing director oversees, the size of the team a partner runs, the depth of the support stack underneath an executive — these are status claims first and economic outputs only incidentally. The private equity acquirer that fires 30% of headcount and finds the firm runs about the same has just performed the empirical demonstration.

The spiritual violence

The deepest part of Bullshit Jobs is the phenomenological one. Graeber argues that work in modern societies has been culturally loaded with the function of meaning-production — you are what you do, your contribution is your worth, the dignity of work is sacred — and that the bullshit job inflicts spiritual violence on the worker by paying them well to do the opposite of this. The worker has been promised meaning and given theatre. The mortgage is paid. The soul is not.

The testimonies in the book are striking precisely because the workers are not under-paid. Many are paid better than the people doing visibly useful work — the nurses, teachers, care workers, sanitation staff, agricultural labourers, artists. Graeber repeatedly observes the inverse relation in contemporary advanced economies between the social usefulness of work and its compensation. The more obviously the work matters, the worse it tends to be paid; the more dispensable the work, the more lavishly it tends to be rewarded. The book treats this as a structural feature, not a coincidence.

The corrosion is documented in the testimonies. Depression. Drinking. Cynicism. The slow erosion of the worker's confidence that there is meaningful work to be done at all. The hours the institution paid for were also hours of life the worker no longer has. The Codex reads this as a debt the institution incurred — Manuṣya Ṛṇa, the debt to fellow humans — and quietly defaulted on.

The neighbours — and where the diagnosis thickens

Graeber's diagnosis sits next to several others that diagnose the same substrate from different angles. The neighbours to know:

• Ivan Illich, counterproductivity: institutions past a complexity threshold produce the inverse of their stated purpose. Bullshit jobs are the inside-the-firm counterpart of medicine sickening and schools un-educating — the work-producing institution starts producing anti-work.
• Byung-Chul Han, burnout society: the achievement subject exploits itself in pursuit of metrics whose ultimate beneficiary is not them. Bullshit jobs are Han's burnout substrate in its purest form — the work is unnecessary, the worker exhausts themselves performing necessity, the only thing produced is the appearance of having worked.
• Karl Polanyi, fictitious commodities: labour is one of the original fictitious commodities — something not produced for sale but commodified as if it were. The bullshit-job phenomenon is Polanyi's commodification of labour at its most reductive: the firm buys labour-hours regardless of whether labour-hours are needed, because purchasing the time is what hierarchy requires.
• E. F. Schumacher, Buddhist economics: the function of work is threefold — to give a person a chance to utilise and develop their faculties, to enable them to overcome ego-centeredness by joining with others in a common task, and to bring forth the goods and services needed for a becoming existence. By Schumacher's definition, the bullshit job systematically fails all three.
• Joseph Tainter, collapse of complex societies: institutions add layers of complexity until marginal returns go negative. The taskmaster layer is Tainter's complexity premium with the units made vivid; bullshit jobs are the marginal-cost-of-complexity made payroll.
• Hannah Arendt, The Human Condition: the distinction between labour (recurring life-maintenance), work (the making of durable artefacts), and action (politics, meaningful public speech). The bullshit job belongs to none of these — it is a fourth category Arendt did not have to theorise, the simulation of work that fulfils the function of neither labour nor work nor action.

The Conway-Debt reading — bullshit jobs as managerial gliders

The Conway Debt frame reads bullshit jobs as a specific kind of institutional skeuomorph. Every layer of managerial hierarchy that ever existed in a large organisation left an artefact — a reporting line, a headcount allocation, a budget category, a job title. The artefacts outlast the conditions that justified them. New managers inherit the artefacts and rebuild around them. The pattern propagates — the manager who took over a team of eight does not reduce the team to four even when four would suffice, because team-of-eight is the artefact they inherited, and the artefact is the basis on which their own status is calibrated. The glider does not stop. The bullshit jobs compound.

This is also why bullshit jobs survive cost-cutting exercises with such striking resilience. The cuts hit the duct-tapers and the visibly-useful junior layers first, because those are the layers least defended by the managerial-feudal logic. The taskmasters, flunkies, and box-tickers persist — because they are the substrate of the hierarchy, not its periphery.

The Indic counter-frame — dharma, Manuṣya Ṛṇa, puruṣārtha

The Pañca Ṛṇa frame answers Graeber at the level of the substrate. Manuṣya Ṛṇa — the debt to fellow humans — names the worker's life-hours as a load-bearing obligation, not a commodified input. The institution that purchases hours without putting them to use is in debt to the human whose hours it has consumed. The accounting surfaces the obligation; the surfacing forces the question of how it will be discharged. A workplace operating under Manuṣya Ṛṇa accounting cannot, under the explicit terms of its own ledger, sustain bullshit work without recording the debt.

The puruṣārtha frame — the fourfold of dharma, artha, kāma, mokṣa — does the same at the level of personal life. Artha is wealth, livelihood, the means of worldly support; it is one of the four, not the totality. Dharma is the orienting purpose under which artha is pursued. The bullshit job is the precise structural consequence of pursuing artha decoupled from dharma — work that pays but does not bear sense. The puruṣārtha frame is Goodhart-resistant by construction; the four hold each other in tension and no single one can be optimised in isolation without producing the failure mode the other three were holding off.

The Techno-Memetic Commons licence and the federated-unicorn architecture do the structural work. A federated proprietor- run cell, accountable to its commons and to its members rather than to an unrelated shareholder, has no upstream incentive to accumulate the hierarchy that produces bullshit jobs in the managerial-feudal firm. The pyramid runs flat. The status hierarchy gets compressed into work hierarchy. The work hierarchy gets stress-tested by membership rather than by quarterly reporting. The bullshit-job substrate becomes structurally expensive.

This is not a claim that the bullshit job disappears under a different substrate. It is a claim that the structural incentive to manufacture it disappears, and the resilience of the bullshit-job pattern under cost pressure becomes a weakness instead of a strength.

What to do with this

Three operating heuristics for builders, founders, and allocators in 2026:

1. Ledger the hours, not just the headcount. If your organisation cannot articulate what each role contributes that someone outside the firm would pay for, or what civic, ecological, or social obligation it discharges, the role is a candidate for the bullshit-job audit. Run the audit privately, with the role-holders, against their own judgement. The accuracy of self-report on this question is, empirically, high.
2. Compress the hierarchy. The managerial-feudal accretion pattern requires depth — the taskmaster layer needs underlings, who need duct-tapers, who need box-tickers. Federated, cooperative, proprietor-run, partnership, and commons-licensed organisational forms have structurally fewer rungs; federated unicorn is one specific architecture for keeping the pyramid flat at scale.
3. Reward usefulness, not visibility. The inverse relation between social usefulness and compensation is a substrate feature, not a market outcome. Inside any single organisation it can be reversed by deliberate policy: pay scales calibrated to demonstrated contribution to the Pañca Ṛṇa ledger rather than to managerial-feudal status display. It is harder than it sounds, but the harder thing is the right thing.