The question most readers bring to this subject is disarmingly practical. They have a company in mind, perhaps their own employer, perhaps a firm whose annual report they have just read, and they want to know how much surplus value that company takes and how the taking works inside an organisation that employs software engineers, franchise managers, warehouse pickers, designers, drivers, and sales staff rather than the mill hands of a nineteenth-century engraving. The Marxian framework was built on the factory, and a fair suspicion follows it into every discussion of the modern firm: perhaps the categories simply do not survive the journey from the spinning mill to the software campus, the logistics platform, and the outsourced supply chain. This article tests that suspicion by doing the work rather than gesturing at it. It takes four kinds of firm, applies the same set of questions to each, and reports where the framework fits cleanly, where it strains, and where it honestly cannot reach.

One conclusion should be announced at the start, because everything else in the article hangs on it and because it disappoints the most common hope readers bring. You cannot compute a company’s rate of exploitation from its published accounts. That is not a complaint about missing data or a plea for better disclosure. It is a structural fact about how accounts are built, and this article will name it the accounts warning: no rate of exploitation can be read off a published income statement, because the accounts are constructed on categories that cut across the Marxian ones at several separate points, so any figure produced that way measures the accounting convention rather than the relation it claims to describe. The popular alternative, dividing a firm’s profit by its wage bill and calling the quotient a rate of exploitation, is wrong for at least four identifiable reasons, and each reason will be examined in turn.

What can be done, and what this article does instead, is more modest and more useful. A Marxian analysis of a firm can identify where value is produced inside and around it, which workers count as productive labour in the technical sense, where surplus value is realised in money, which incomes are shares of surplus value produced elsewhere, and which mechanisms move the rate of surplus value up or down. It can say a great deal about the structure of a business model. It can explain why a loss-making firm may still organise the extraction of surplus value and why a highly profitable firm may produce very little of its own. What it cannot do is hand the reader a single defensible percentage for a named company, and any source that offers one without a long account of its reclassification decisions is selling false precision. The foundation for everything here is the concept itself, set out in full in the surplus value complete guide, and this article assumes that foundation while concentrating on the application.

The firm as a circuit, not a pile of money

A modern company is best approached, in Marxian terms, not as a stock of assets or a stream of earnings but as a circuit that money capital travels again and again. Money is advanced to buy means of production and labour power. Production follows, in the course of which labour power is consumed and new value is created, part of it replacing the value of the labour power itself and part of it exceeding that value. The product is then sold, and the value it carries, including the surplus portion, is realised in money. The enlarged sum is advanced again. Every department of a large firm, and every puzzle this article examines, belongs somewhere on that circuit, and most confusion about surplus value in firms comes from mistaking one station on the circuit for the whole journey.

The distinction between producing value and realising it does heavy work throughout. A commodity that is made but never sold embodies labour that the market refuses to validate, and the surplus value aimed at in its production is never realised. A commodity that is sold at a high price may realise more value than was produced in making it, because prices redistribute value among producers as well as registering it. Production and realisation are both necessary to the firm, and the workers who perform them are all necessary to the firm, but necessity to the firm is not the criterion Marx uses when he sorts labour into productive and unproductive. That sorting, which looks pedantic from a distance, turns out to be the hinge on which the analysis of retail, finance, platforms, and management all swing.

A second piece of ground clearing concerns scale. The rate of surplus value, surplus value divided by variable capital, is a ratio defined for a mass of capital employing labour in production. It is not defined for a brand, a logo, a patent portfolio, or a stock market valuation, and attempts to attach it to those things produce nonsense even when the arithmetic is careful. Firms, moreover, are legal shells around activities that may span several countries and several distinct circuits at once. A single corporate group may run factories that produce value, shops that realise it, a finance arm that lends against it, and a property arm that collects rent from it. The accounts consolidate all of that into one set of lines. The Marxian analysis has to disaggregate first, and disaggregation is a theoretical act, not a clerical one, because the categories being separated are not printed anywhere in the source documents.

Third, the analysis is structural rather than biographical. To say that surplus value is extracted in a workplace is not to say that a manager is cruel, a founder is greedy, or a shareholder is wicked. The extraction follows from the wage relation itself, from the fact that labour power is bought for a period and its use creates more value than its own reproduction costs. Sympathetic employers extract it and ruthless ones extract it, at rates set by technology, organisation, the length and intensity of the working day, and the value of the goods workers consume. Readers who want a verdict on individual conduct will find this article cold, and the coldness is deliberate. A framework that can only describe firms run by villains would be useless for analysing firms run by anyone else.

Surplus value inside the firm: the mechanism restated for workplaces that do not look like factories

The mechanism itself does not change when the workplace changes. An employer advances a wage that represents the value of labour power, meaning the value of the goods and services required to maintain the worker and raise the next generation of workers at a historically given standard of life. The employer then directs the worker’s activity for the contracted period. During part of that period the worker produces value equal to the wage. During the remainder the worker produces value for which no equivalent is returned. That remainder, in labour time, is surplus labour, and its product, in value terms, is surplus value. Nothing in this account requires the workplace to contain a chimney, a whistle, or a loom. It requires a wage bargain, a period of directed labour, and a product or service that is sold.

Consider how little of the mechanism depends on the physical character of the output. A machinist cutting metal and a programmer writing code both sell labour power rather than a finished product, both work under direction for a set period, and both produce something the employer owns and sells. The machinist’s output can be weighed. The programmer’s output can be copied at almost no cost, which raises hard questions about price and rent that a later section takes up, but the labour relation is the same relation. Marx’s own examples already included teachers in privately owned schools, singers employed by an impresario, and the staff of commercial houses, because he understood the category to follow the social relation rather than the material of the product. The modern firm multiplies the variety of materials without altering the underlying form.

What does change, from firm to firm, is everything around the mechanism. The proportion of the workforce engaged directly in production varies enormously. The distance between the place of production and the place of sale varies. The share of the surplus claimed by lenders, landlords, licensors, and the state varies. The degree to which production is subcontracted to other firms, other countries, or nominally independent workers varies. These variations are exactly what make firm-level analysis interesting and exactly what defeat the attempt to read a rate of exploitation off consolidated accounts. The mechanism is uniform. Its clothing is not, and the accounts describe the clothing in categories designed for investors and tax authorities, not for value theory.

There is also a temporal point that modern workplaces make vivid. Labour power is increasingly bought in fragments: by the hour, by the shift, by the delivery, by the completed task, by the month of a subscription contract for services. Fragmentation does not abolish the division of the working period into necessary and surplus portions. It conceals the division more effectively, because a worker paid per task cannot easily see which tasks repaid the cost of their own labour power and which were performed beyond that point. Piece payment was already, in Marx’s discussion, a form that made the wage look like a price for output rather than a price for labour power. Task-based payment across a digital platform extends that appearance to employments where no factory, no foreman, and no shared workplace exist at all.

Why accounting profit is not surplus value

Can you calculate a company’s rate of exploitation from its accounts?

Not reliably. Published accounts mix productive and unproductive labour in one wage line, book profit where it is taxed rather than where value was produced, and net out interest and rent that are shares of surplus value rather than deductions from it. The categories do not align, and no adjustment printed in the notes repairs the misalignment.

The calculation that circulates online takes a company’s reported profit, divides it by the reported wage bill, and presents the result as the rate of exploitation. It has the great virtue of being performable by anyone with an annual report and a spreadsheet, and the great defect of being wrong in at least four separate ways at once. Naming those ways precisely is the most useful service this article can perform, because each one corresponds to a real feature of how firms and accounts work, and because together they constitute the accounts warning in its full form.

The first reason is the composition of the wage line. A company’s accounts gather every salary it pays into broad staff cost categories: the machinist and the marketing manager, the software engineer and the internal auditor, the delivery driver and the director of corporate communications. In the Marxian classification these are not the same kind of expenditure. Some of those wages are variable capital, advanced to buy labour power that produces value and surplus value. Others are paid out of surplus value already produced, as a cost of circulation, supervision, compliance, or administration. Accounting profit is therefore net of the wages of unproductive workers, whose pay is a deduction from surplus value rather than a cost of producing it. Dividing profit by the whole wage bill compares a residue with a denominator that mixes two different economic categories, and the result belongs to neither.

The second reason is that reported profit excludes surplus value transferred out of the firm before the profit line is reached. Interest paid to lenders is, in the Marxian scheme, a share of surplus value, a portion of the unpaid labour of the productive workforce claimed by money capital. Rent paid for land and premises is likewise a share, claimed by property. Royalties and licence fees paid for intellectual property are, on the analysis offered later in this article, largely shares of the same kind. A firm’s operating and net profit figures are calculated after many of these transfers, or present them in lines that the quick calculation ignores. The surplus value produced by a firm’s productive workers is therefore systematically larger than the profit the firm reports keeping, by an amount that varies with how the firm is financed and housed. Two firms with identical production and identical workforces, one owning its premises outright and funded by equity, the other renting and funded by debt, will show different profits and identical surplus value extraction. The quick calculation ranks the indebted tenant firm as less exploitative. The ranking is an artefact of its financing structure.

The third reason is geography, and it has grown more important as production has been organised across borders. Profit is booked in the jurisdiction where the firm chooses to recognise it, which is a decision shaped by tax law, transfer pricing, and the location of intangible assets, rather than in the place where value was produced. A corporate group can arrange for its manufacturing subsidiary in one country to sell output at a modest markup to a distribution affiliate in a low-tax jurisdiction, which then books the large margin on the final sale. The consolidated accounts may be perfectly legal and perfectly audited, and they locate profit where the group’s tax advisers prefer it to sit. Value, on the Marxian account, was produced where the labour was performed. The accounts say nothing reliable about that location, and a rate of exploitation computed for the group as a whole dissolves the very differences in wages and conditions between sites that a serious analysis wants to expose.

The fourth reason concerns depreciation and the treatment of constant capital. Accounts write down machinery, buildings, and software according to depreciation schedules set by accounting standards and tax rules. These schedules are conventions. They determine how much of the value of fixed assets is charged against each period’s revenue, and they move reported profit up and down without any change in production. Marxian value accounting has its own treatment, in which fixed capital transfers its value to the product piecemeal across its working life, but the transfer period is the asset’s actual service life in production, not a schedule chosen within permitted bands. A firm can lengthen a depreciation schedule, raise reported profit, and alter nothing about the labour performed in its workplaces. The quick calculation dutifully reports a rise in exploitation. Nothing of the kind has occurred.

Behind these four reasons sits a deeper one, which the later article on measuring the rate of exploitation develops for national economies. Even a perfect set of firm accounts would record prices, and surplus value is a category of value. The transformation of values into prices of production redistributes surplus value among firms according to the size of their total capital, so that a capital-intensive firm systematically receives more profit than the surplus value its own workers produce, and a labour-intensive firm systematically receives less. This is not an accounting defect. It is how the system distributes the pool, and it means the gap between a firm’s profit and its own workers’ surplus value is not noise around a true figure but a designed feature of capitalist competition. The full separation of the two concepts is the subject of the comparison with surplus value and profit, and the reader who keeps only one sentence from this section should keep this one: profit is a firm’s share of a social pool, while surplus value is what a workforce produces, and shares and products coincide only by accident.

Productive and unproductive labour: the distinction that sorts the modern workforce

What makes labour productive in the capitalist sense?

Labour is productive in the capitalist sense when it is exchanged against capital and produces surplus value, regardless of whether its output is useful, tangible, or admirable. A teacher in a fee-charging school is productive for the school’s owner. A teacher in a public school, doing identical work, is not, because no capital is valorised by it.

Chapter Sixteen of Capital Volume One is titled to announce that the distinction has two senses, and the capitalist sense is the narrower and stranger one. In the general sense, productive labour is labour that produces a use value, that makes or does something. In the capitalist sense, productive labour is labour that produces surplus value. The two senses overlap often and diverge sharply at the edges, and the divergences are the analytically interesting part. A cook employed by a restaurant chain is productive labour for the chain. A cook employed by a household, preparing the same meals with the same skill, is unproductive labour for the household, because the household spends revenue rather than advancing capital and expects dinner rather than an enlarged sum of money. Nothing about the cooking differs. Everything about the social relation differs, and Marx’s category tracks the relation.

This is why the distinction so often offends first-time readers. It appears to rank workers, placing the factory hand in an honoured class and the nurse, the civil servant, and the carer outside it. It ranks nothing of the kind. It classifies expenditures. To call labour unproductive in the capitalist sense is to say that its wage is paid out of revenue or out of surplus value produced elsewhere, not that the work is useless, easy, or socially unnecessary. Some of the most indispensable labour in any society is unproductive in this technical sense, and some thoroughly trivial labour, manufacturing novelties nobody needs, is productive because a capitalist sells the novelties at a profit. Marx was aware of the provocation and leaned into it, because the provocation exposes something real: capitalism does not ask whether work is needed before it asks whether work valorises capital, and the gap between those questions is one of the system’s defining tensions.

Inside a modern firm the distinction cuts through nearly every department. Production workers, whether they shape materials or write the software that is the product, are productive. Workers in pure circulation, buying and selling, handling money, keeping the commercial books, and marketing finished goods, are unproductive in the strict sense, because circulation transfers ownership of value without adding to it, a point Volume Two of Capital establishes through the analysis of the circuits. Workers in transport and storage occupy a middle position that Marx treats as productive, because moving a commodity to where it can be used and preserving it until it is used continue the production process into the sphere of circulation. Workers in maintenance and repair of the firm’s own equipment are productive insofar as they preserve the value of constant capital in use. Workers in supervision divide, as a later section explains, between coordination that any complex production requires and control that the antagonism between capital and labour specifically demands.

The classification matters for the accounts warning because it cannot be read off payroll data. Job titles do not settle it. A firm may call every employee a team member, a partner, an associate, or a creator, and the vocabulary of human resources departments is designed for morale rather than analysis. The same titled role can straddle the line: a designer who creates the product is productive, while a designer who creates the advertising campaign for the product performs circulation labour. A software engineer building the application that subscribers pay for is productive, while an engineer building internal reporting tools serves administration. Any serious firm-level analysis must therefore work from what people actually do, task by task, and must admit that a substantial share of a modern workforce’s hours resist clean classification. That admission is not a failure of nerve. It is the honest condition of applying a nineteenth-century distinction to integrated white-collar work, and the objection section later in this article gives the strongest version of the argument that the resistance is fatal.

One more refinement prevents a common error. Unproductive labour is not paid for nothing, and unproductive workers are not parasites on the productive ones in any moral sense. Circulation workers perform real and necessary functions: commodities must be sold, accounts kept, payments processed, and legal obligations met, or the circuit stops. Their labour is exploited in the broader sense that they work longer than the labour time their wages represent, a point Marx makes about commercial employees specifically, since their unpaid labour reduces the deduction their employment makes from surplus value even though it creates none. A commercial worker whose working day exceeds the portion covered by the wage performs surplus labour that lowers the costs of realisation for the employer. The worker produces no surplus value and is nevertheless worked beyond the necessary portion of the day. Keeping those two claims together, apparently contradictory and in fact compatible, is a good test of whether a reader has grasped the distinction or merely memorised a slogan about it.

The four-firm analysis grid

Arguments about surplus value and firms go astray most often by generalising from one kind of business. The remedy used throughout this article is a fixed set of questions, applied in the same order to four firm types, so that the differences in the answers reflect differences in the firms rather than in the analyst’s mood. The four-firm analysis grid below states the results compactly. The sections that follow it argue each row in prose, and readers who disagree with a cell should be able to locate the disagreement in a specific claim rather than in a general impression of the article’s politics.

Firm type	Where value is produced	Who counts as productive labour	Where surplus is realised	Principal analytical difficulty
Manufacturer	In the plant, where labour transforms materials into a finished commodity	Production operatives, technicians, maintenance and internal transport staff	On sale of the finished commodity to distributors or final buyers	Separating the group’s production sites from its sales, finance and property arms in consolidated accounts
Software company	In the labour of building and maintaining the application or service that is sold	Engineers, designers and testers whose work constitutes the product being sold	Through licence fees, subscriptions and service contracts paid by customers	Near zero reproduction cost makes price rest heavily on intellectual property and market position rather than on value produced
Retailer	Value is not produced in the shop; it is produced upstream by suppliers’ workers	Shop staff are unproductive in the strict sense, while delivery and warehouse transport labour is productive	In the act of sale itself, where the retailer takes a commercial share of surplus value produced elsewhere	Explaining a large and visible profit that corresponds to no value produced on the premises
Platform	In the labour coordinated through the application, such as driving, delivering or hosting	Drivers, riders and hosts, on the analysis that treats the relation as disguised wage labour	Through the commission or fee deducted from each transaction before the worker is paid	The worker owns part of the means of production, which strains the standard account of the wage relation

The grid’s last column deserves emphasis, because it is where this article parts company with confident popular treatments. Each firm type has a characteristic point at which the framework creaks. For the manufacturer the creak is bookkeeping: the production is legible but the accounts fuse it with everything else the group does. For the software company the creak is price formation: the product can be reproduced almost freely, so the revenue stream behaves like a toll on access rather than a payment for embodied labour. For the retailer the creak is explanatory: the profit is real, large, and historically durable, yet the theory insists no value is produced on the shop floor, so the profit must be traced to a share in value produced by other people’s workers in other people’s firms. For the platform the creak is the labour relation itself: the worker supplies a significant productive asset, and the theory must decide whether that fact changes the classification of the whole arrangement or merely its costs. Naming the creak in each case is more informative than pretending the framework glides over all four.

The manufacturer: the classical case wearing modern clothing

The manufacturing firm remains the easiest case, and it is worth extracting every lesson it offers before moving to harder ones. Imagine a mid-sized producer of domestic appliances, with two plants, a head office, a sales division, a service network that repairs its machines under warranty, and a finance scheme that lets retailers offer customers instalment terms. Even this unglamorous firm is a composite. The plants produce value and surplus value. The service network divides: repairs that maintain machines still in use conserve value and support realisation, while the administrative apparatus that schedules and bills for them belongs to circulation. The sales division realises value and produces none. The finance scheme collects interest, a share of surplus value, from the incomes of customers and the margins of retailers. A single consolidated profit figure sits on top of all five activities like a lid on a compartmented box.

Within the plants, the familiar levers operate. Lengthening the working day or adding a shift raises absolute surplus value, subject to law, custom, and the physical limits of the workforce. Raising productivity through new machinery and reorganised flow raises relative surplus value, because the same reasoning that applies to the economy applies inside the division of the day: the portion of the shift needed to reproduce the value of the wage shrinks when the goods workers buy grow cheaper, and it shrinks for this firm’s workers through changes in distant industries they never see. Intensification, a faster line, tighter staffing, fewer pauses, raises the expenditure of labour per hour and functions as a hybrid that managers pursue precisely because it is less visible than a longer day. None of these levers requires the firm’s owners to think in Marxian vocabulary. The vocabulary describes what the levers do, whoever pulls them and whatever they call them.

The manufacturer also displays, in miniature, the difference between producing surplus value and keeping it. Suppose the firm sells mainly through a dominant retail chain. The chain’s buying power lets it dictate prices that compress the manufacturer’s margin, transferring part of the surplus value produced in the plants to the retailer as commercial profit. Suppose the firm rents its newer plant from a property company; rent takes another share. Suppose it borrows to finance the machinery; interest takes a third. The residual profit in the manufacturer’s accounts may look modest beside the wage bill, and the quick calculation may declare exploitation mild. The workers’ surplus labour, meanwhile, has fed four different incomes, only one of which appears as this firm’s profit. Modesty of profit is compatible with a high rate of surplus value whenever the surplus is widely shared out among capitals and property owners positioned along the circuit.

Automation inside the plant poses the question that a dedicated section below treats generally, but the manufacturer shows its concrete form. A new line replaces forty operatives with six technicians and a great deal of expensive equipment. The firm’s wage bill for production falls sharply. The surplus value produced in this plant falls too, other things equal, because fewer workers perform surplus labour, while the value transferred from the machinery to each unit of output rises. Yet the firm may grow richer, because its lower costs let it sell below the prevailing price, capture market share, and receive through prices a redistribution of surplus value produced across the industry. A single firm’s fortunes under automation therefore prove nothing, by themselves, about exploitation rising or falling. The firm has changed its position in the distribution of a pool whose total depends on the labour employed across the whole branch of production, and confusing the firm’s improved position with an increase in what its own remaining workers produce is the characteristic error of company-level commentary.

The software company: value without a physical product

Do software engineers produce surplus value?

Yes, when they are employed by a capital to build a product or service that is sold, because the criterion is the relation and not the material of the output. Code that is sold, licensed, or supplied as a service embodies labour in the relevant sense, and the engineer’s working period divides into necessary and surplus portions like any other.

Scepticism about software usually rests on a picture of value as something that must be baked into a physical object, like moisture into bread. The picture is wrong even about bread, since value for Marx is a social relation expressed in exchange rather than a physical ingredient, and it collapses entirely for services, performances, transport, and tuition, all of which he treated as capable of bearing value. A subscription application is a use value delivered continuously: storage held available, computation performed, records kept, interfaces maintained. The labour that builds and maintains it is expended under the direction of an employer, for a wage, across a working period that divides into the portion replacing the wage and the portion beyond it. The division is less visible than on an assembly line because the product is intangible and the working day is often loosely bounded, but visibility is not a theoretical category.

The genuine difficulty arrives with reproduction. A physical commodity must be produced afresh for each buyer, so the labour embodied in the stock sold each period corresponds roughly to the labour performed each period. A software product is produced once and reproduced at a cost approaching zero. The thousandth copy requires no thousandth portion of the original development labour. Revenue, however, does not approach zero, because access is fenced by law and by technical control. The price of the copy is therefore sustained not by the labour needed to reproduce it but by the legal monopoly over its use, and the income stream takes on the character of rent on an intellectual asset as much as profit on production. The rent versus profit debate owns that distinction in full, and this article flags it as the software case’s characteristic strain: a Marxian analysis that ignored it would describe subscription revenue as though each payment corresponded to fresh production, which would badly overstate the value produced in the period while understating nothing about the labour relation of the engineers themselves.

Two kinds of software labour should be separated before the strain is exaggerated. Labour that maintains, updates, secures, and extends a living service is continuously productive of the use value subscribers pay for, since the service would decay into uselessness without it; a neglected platform loses data, compatibility, and trust within a short span. Labour that produced the original codebase is better understood on the analogy of fixed capital: its value, such as it is, is transferred across the service’s life rather than recreated in each transaction, and the analogy explains why firms amortise development costs in ways that vaguely shadow the theory while following accounting rules of their own. Neither treatment requires the analyst to pretend that copying is production. Copying is distribution. Distribution by electronic means happens to be nearly free, which is precisely why the industry’s economics hinge on fences, defaults, switching costs, and network effects rather than on the cost of supply.

The workforce composition of a software firm also tests the productive and unproductive distinction more severely than a factory does. Alongside the engineers stand sales staff who court enterprise clients across long cycles, marketers who manufacture demand, customer success staff who keep subscribers from leaving, recruiters in a permanent contest for scarce engineers, and layers of management coordinating work whose output cannot be inspected by walking a floor. A defensible classification exists for each group, and earlier sections supply the criteria, but the honest finding is that the boundary runs through teams rather than between them. An engineer who spends a month building a feature and a month building internal dashboards has performed both kinds of labour in one quarter on one salary. Firm-level estimates that assign whole salaries to one category import an error larger than many of the effects they claim to measure, which is one more reason the accounts warning holds with special force in this sector.

None of this makes the software firm an exception to the theory of surplus value. It makes it a case where the theory’s distinctions must be applied at a finer grain than the firm’s own reporting, and where a large share of the income is plausibly rent-like rather than a direct product of the period’s labour. Engineers are exploited in the technical sense when their labour produces more value than their wages represent, and high salaries do not settle the question either way, since exploitation is a ratio between value produced and the value of labour power, not a judgement about whether a salary feels generous. A scarce specialist may command a wage far above the cost of reproducing labour power in general, yet still work a substantial surplus portion of each week for the employer. Conversely, a firm burning investors’ money on salaries while producing a service nobody buys may be extracting no surplus value at all, because the value its staff aim to produce is never validated in sale. The software sector supplies specimens of both, sometimes within the same company’s history.

The retailer: profit where no value is produced

Retail concentrates the paradox that casual readers find hardest to accept. A large retailer earns large profits, employs enormous numbers of people, and on the Marxian account produces no value at all in its shops. The resolution lies in the division of the total surplus value among different kinds of capital. Commercial capital performs the function of selling, which industrial capital would otherwise have to perform for itself, and it receives a share of the total surplus value in proportion to its size, as commercial profit. The retailer’s margin between buying price and selling price is that share. It is drawn from value produced by the workers of the firms that made the goods, the firms that grew or mined their materials, and the transport workers who moved them. The shop is where a portion of other people’s surplus labour changes hands for the last time before the commodity leaves the circuit.

Shop workers, on this account, are unproductive labour in the strict sense, and the classification says nothing about their effort, their skill, or the unpleasantness of their working conditions, which are frequently worse than those in the factories upstream. Their employer nevertheless extracts surplus labour from them in the form described earlier: the working day of a commercial employee exceeds the portion whose value the wage replaces, and the excess reduces the cost of realisation that the retailer must deduct from its commercial share. A retailer that staffs tills thinly, schedules workers in fragments that follow footfall by the quarter hour, and leaves a single employee to cover a department is lengthening the surplus portion of commercial labour and pocketing the saving. The worker produces no value and is exploited in the everyday sense of being worked beyond the paid equivalent, while the technical statement that no surplus value is produced on the premises remains true. Readers who treat those sentences as contradictory have not yet separated the production of value from its appropriation, which is the separation this whole article exists to teach.

The boundary cases inside retail are instructive. A supermarket that bakes bread on the premises, butchers meat, and prepares meals employs productive labour for those activities, since production is occurring under its roof and direction even though the same building sells soap. Its warehouse and delivery operations employ transport labour, productive on Marx’s treatment of transport as a continuation of production. Its online operation, in which pickers assemble orders in a dark store, blends production-like handling with pure circulation in proportions that defy any quick verdict from outside. Large retailers have also moved steadily upstream, commissioning own-brand goods from contracted manufacturers and capturing, through their buying terms, much of the surplus value those manufacturers’ workers produce. The retailer’s profit then includes both a commercial share and a slice of industrial surplus extracted by bargaining power rather than by ownership, a reminder that the channels through which surplus value travels are set by market position as much as by legal title.

Why does the distinction matter, if the retailer’s profit arrives either way? Because it changes what analysis predicts. If retail profit were produced in the shop, then squeezing shop labour would be the main lever on it, and the retailer’s relentless pressure on suppliers’ prices would be a secondary detail. If retail profit is a share of a pool produced elsewhere, then buying power, logistics, and the terms imposed on suppliers are central, and the condition of the garment worker or the food processing worker at the far end of the chain belongs inside the analysis of the retailer’s accounts. The second picture directs attention to the right workplaces. It also explains a pattern that puzzles commentators who assume profit must be made where it appears: retail margins can be squeezed by competition among retailers even while the exploitation of the upstream workforce intensifies, because the pool is divided by bargaining among capitals after it is produced by labour.

The platform: the puzzle of the worker who owns part of the means of production

The platform firm, matching riders, drivers, hosts, couriers, or freelance workers with customers through an application it owns, presents the hardest classification problem in this article, and the brief for it is right to call the case a genuine theoretical puzzle rather than a debating point. The standard account of surplus value presupposes a worker separated from the means of production, who therefore sells labour power to an owner. The driver who owns a car, the host who owns a flat, the courier who owns a bicycle and a phone appear to break the presupposition. They bring a productive asset to the arrangement. Are they workers, small businesses, or something the old categories cannot hold?

Begin with what the platform owns and does. It owns the application, the matching system, the payment channel, the customer relationship, the rating apparatus, and the brand under which demand is gathered. It sets the price of the service in most such markets, allocates work by an algorithm the worker cannot inspect, disciplines through ratings and the threat of deactivation, and deducts a commission from every transaction before the worker is paid. The worker owns the vehicle and bears its costs: fuel, insurance, maintenance, depreciation, and the dead time between paid tasks. Ownership, examined closely, is split, and the split is not even. The platform owns the means of access to the market, without which the car is merely a car. The worker owns the means of performing the task, without which the platform has nothing to sell. Each side holds an asset the other needs, and the commission is the price at which their claims are settled transaction by transaction.

Two serious Marxian answers compete, and this article presents both before judging between them. The first treats the platform relation as a disguised wage relation. On this reading, the decisive facts are control and dependence: the platform fixes the price, directs the work through allocation and ratings, monopolises the customer connection, and can expel the worker unilaterally, while the worker depends on the platform for the bulk of an income and cannot realistically take the customer base elsewhere. The car, on this reading, is analogous to the tools some historical workers supplied within an unmistakably subordinate relation, or to the instruments a home-based pieceworker owned under the putting-out system, where merchant capital controlled the market and the worker’s ownership of a loom settled nothing about who appropriated the surplus. The commission then functions as the form in which surplus value is extracted, and the worker’s bearing of vehicle costs functions as a transfer of constant capital expenses from the firm to the worker, raising the extracted share. The fullest treatment of this reading is given in the analysis of platform capitalism, which this article follows on the facts of control.

The second answer treats the driver as a petty commodity producer, a small independent business selling a service in a market, and treats the commission as rent: a payment for access to a scarce and monopolised means of circulation, comparable to the rent a market trader pays for a pitch or the fee a franchisee pays for a brand and a territory. On this reading the driver is not exploited in the technical sense by the platform, because no labour power has been sold. The driver may still be squeezed, dominated, and impoverished by the relation, and may work brutally long hours in self-exploitation, driving the value of their own labour power below its normal level in order to survive the fee structure. The platform’s income, meanwhile, is a deduction from the value the driver produces, extracted by monopoly of access rather than by the wage relation. This reading takes the property facts seriously and refuses to dissolve them into a metaphor about disguise.

The verdict this article defends is that the two readings divide the sector rather than one refuting the other, and that the dividing line is the degree of control in the specific arrangement. Where the platform sets prices, allocates tasks, penalises refusal, and prevents the worker from building an independent clientele, the wage relation exists in substance and the disguise reading is correct, whatever the contract calls the parties. Where the worker sets prices, chooses clients, works through several channels, and can leave with a reputation and a customer list intact, the petty production reading fits, and the commission is rent on access. Many real arrangements sit between the poles, and they sit there because platforms calibrate control carefully, seeking the direction of a workforce without the obligations of an employer. A calibration designed to fall between legal categories will also fall between theoretical ones, and the analyst’s task is to measure the control actually exercised rather than to ask which box the business model prefers. What neither reading permits is the lazy conclusion that owning a car settles the question. Ownership of one asset, inside a relation structured by another party’s ownership of the market itself, has never been the criterion in Marx’s work, where the putting-out worker with a loom and the peasant with a plot both appear in relations of subordination to capital that owned no part of their tools.

A loss-making platform sharpens the lesson. Such a firm may spend years subsidising rides and deliveries with investors’ funds, reporting losses while its commissioned workers perform surplus labour on every trip. The surplus value embodied in the service is realised at prices deliberately set below its value, which means part of it is transferred to customers as cheap journeys and part absorbed in building the network’s dominance, while the firm’s accounts show a deficit financed from capital raised on the expectation of future monopoly rents. Extraction, realisation, and reported profit come apart completely. An analyst who read the loss as proof that no exploitation occurred would have the matter exactly backwards: the workers’ surplus labour is one of the assets being spent, alongside the investors’ cash, to purchase a market position from which rents and surplus value can later be drawn together.

Global supply chains: where value is produced, where it is booked, and where it is captured

The standard illustration in the global value chain literature involves a consumer electronics product whose components are made in several countries, whose assembly takes place in another, whose design and branding sit in a further one, and whose profit is recognised in a jurisdiction chosen for its tax treatment. This article follows the instruction of its brief and uses the structure without attaching figures to any named company, since the structure is what carries the analysis and since invented or half-remembered corporate numbers would violate the standards of the series. Walk the chain slowly. Minerals are extracted and processed. Components are fabricated in plants where wages are a small fraction of those in the brand’s home market. Assembly gathers the components through the labour of workers whose working days are long and whose necessary labour is short, because the value of labour power where they live is set by a local standard of life far below that of the markets where the finished product will sell. The product is shipped, marketed under a powerful brand, and sold at a price that bears no visible relation to the assembly wage.

Where is the value created? On the Marxian account, at every point where productive labour is performed under capital: in the mine, the component plant, the assembly hall, and the transport links, in proportion to the socially necessary labour expended at each. Where is the surplus value realised and captured? Disproportionately at the ends of the chain that the literature calls the high-value segments, design, branding, logistics control, and retail, which are precisely the segments where the least value is produced relative to the income received. The brand captures value made in other countries through several mechanisms working together. Its intellectual property commands licence-like payments inside its own group. Its buying power sets component and assembly prices that leave suppliers with thin industrial profit while the brand’s margin stays wide. Its control of the customer relationship lets it appropriate, as commercial profit and as rent on the brand itself, a share of the pool far exceeding the surplus produced by its own employees. The chain thus operates as a machine for producing surplus value where labour is cheap and appropriating it where property rights and market access are concentrated, and the wider treatment of that machine belongs to the guide to imperialism and global supply chains.

Two cautions keep this analysis honest. The first is that cheap labour is not automatically more exploited labour in the technical sense. The rate of surplus value depends on the division of the working day, and a low-wage workforce with low productivity may perform a smaller surplus portion than a high-wage, high-productivity workforce whose output per hour is many times greater. Assembly workers in low-wage sites are frequently exploited at high rates because their working days are long and the value of their labour power is low, but the claim must be argued from day length, intensity, and productivity in each case, not inferred from the wage level alone. Commentators who equate a low wage with a high rate of exploitation commit the same error as those who equate a high salary with freedom from exploitation, and both errors dissolve the ratio into a feeling about pay. The second caution is that value transfer through prices is not a conspiracy and requires no secret meetings. Prices of production, monopoly positions, intellectual property law, and ordinary bargaining between unequal firms do the transferring in the open, inside audited contracts. The Marxian claim is about the structure that makes the transfers systematic, not about hidden misconduct by the firms that benefit from it.

For the analyst examining a single branded firm, the chain perspective reverses the usual reading of the accounts. A brand company may employ relatively few productive workers of its own, report a modest wage bill, and earn very high profits, and the quick calculation then yields an absurd rate of exploitation for its headquarters staff, who are largely engaged in design, marketing, and administration. The absurdity is the tell. The profit is not the surplus value of the headquarters workforce. It is a capture, through property and market position, of surplus value produced by workforces the brand does not employ, in firms it does not own, in countries whose names appear nowhere in its segment reporting. Any framework that cannot say this will misdescribe the most characteristic corporate form of the global economy, and the ability to say it, precisely and without invented figures, is among the strongest arguments for keeping the Marxian categories in working order.

Automation and the inversion of the two rates

Why does automation pull the two famous rates in opposite directions?

Automation raises the machinery set in motion by each worker, which lifts productivity and tends to raise the rate of surplus value by shrinking necessary labour. That change raises constant capital relative to variable capital, which tends to lower the rate of profit. One tendency describes the division of the working day. The other describes the return on capital advanced.

The highly automated factory is the case where popular discussion most reliably inverts the theory. Fewer workers on the floor invites the conclusion that exploitation has faded with the headcount, as though a plant with robots were a plant where the old relation had been engineered away. The Marxian analysis gives a more precise and less comforting answer. Each remaining worker now operates equipment embodying a large mass of past labour, and the output per worker rises steeply. The value of labour power, meanwhile, is pressed downward over time as productivity rises in the industries supplying workers’ consumption, and the remaining workers may also work at greater intensity, monitoring several processes, intervening at speed when systems fail, and carrying responsibility for stoppages whose cost per minute has risen with the value of the line. The rate of surplus value for those workers can therefore be very high. Exploitation, being a ratio, does not shrink when the workforce shrinks. It can intensify per worker while the total mass of surplus value produced in the plant falls.

The rate of profit moves the other way for a structural reason. Profit is measured against the whole capital advanced, and automation swells the constant portion of that capital enormously while shrinking the variable portion, the only portion that expands in production. A plant can thus exhibit a rising rate of surplus value and a falling rate of profit simultaneously, and the firm’s owners experience the second figure, not the first. This divergence is the seed of the tendency examined in the article on the falling rate of profit, and it explains a recurring spectacle in capital-intensive industries: firms that squeeze remarkable surplus labour from small, highly skilled workforces while complaining, sincerely, that returns on their investment are disappointingly thin. Both the squeeze and the complaint can be genuine. They are reports from different ratios.

Judgement about a specific automated workplace still requires care, because automation changes the composition of labour as well as its quantity. The technicians who remain may be scarce and well paid, and part of their salary may represent a share in the firm’s rents rather than the ordinary value of labour power, a possibility the analysis must allow when skills are rare and bargaining power is real. The displaced workers, meanwhile, do not vanish from the theory. They enter the relative surplus population, and their availability disciplines the wages of those still employed, which is one of the channels through which machinery raises the rate of surplus value across an industry beyond the walls of any single plant. Assessing the automated factory in isolation therefore understates its effect twice over: it misses the intensity borne by the workers inside and the pressure exerted on the workers outside. The honest summary is that a highly automated factory is typically more exploitative per remaining worker and less profitable per unit of capital than the labour-intensive plant it replaced, and that anyone who uses its small headcount as evidence of a gentler capitalism has confused the two rates the theory exists to keep apart.

Managers, supervisors, and the labour of direction

Few questions arrive as reliably in discussions of this subject as the one about managers, and few answers benefit as much from a distinction Marx himself drew. The work of directing a combined labour process has, he argued, a double character. Part of it arises from the nature of cooperative production as such: any enterprise in which many people work in coordination needs planning, sequencing, communication, and the resolution of friction, just as an orchestra needs a conductor whose contribution has nothing to do with property relations. Part of it arises specifically from the antagonism between capital and labour: the need to extract effort from workers who have no interest in the surplus their effort creates, to monitor, to discipline, to set pace against resistance. The first part is a function of production. The second is a function of exploitation itself.

On that division, the manager’s salary decomposes conceptually even where no payroll system could separate it. Insofar as a manager performs the coordinating function, the labour is productive in character, part of the collective labourer whose combined activity makes the product. Insofar as a manager performs the control function, the salary is a deduction from surplus value, a cost the system pays to secure the extraction, functionally similar to the wages of other unproductive staff. Senior executives complicate the picture further, because their remuneration frequently includes a share of the surplus itself, in bonuses and equity, which places part of their income on the side of appropriation rather than on the side of wages altogether. A chief executive paid partly in shares is not simply a very expensive employee. That part of the package is a claim on surplus value, exercised from inside the firm, and treating the whole package as a wage cost in a quick exploitation calculation buries the fact.

The practical consequence for firm analysis is humility about layers. A flat organisation with self-managing teams has not abolished the control function; it has distributed it among peers, software dashboards, and customer ratings, where it may operate more continuously than any foreman could. A tall hierarchy has not thereby wasted its salaries; some of those layers perform coordination without which a complex product would disintegrate. What the theory supplies is not a headcount of villains but a question to ask of each managerial activity: would this task still be needed if the workers owned the enterprise and directed it themselves? Coordination survives that question. Surveillance aimed at intensifying unpaid labour does not. The question will not produce a number, and readers by now will recognise the pattern of this article: the framework yields discriminations, mechanisms, and structures, and it refuses, on principle and on evidence, to yield the single percentage that popular treatment keeps demanding.

Outsourcing, franchising, and the moving boundary of the firm

Outsourcing rearranges the extraction of surplus value without reducing it, and the rearrangement is the point. When a firm dismisses its cleaners, caterers, drivers, or production workers and buys the same services from a contractor, its own wage bill falls, its own headcount shrinks, and its accounts present a leaner operation. The labour is still performed, by workers whose employer is now the contractor, and surplus value is extracted from them inside the contractor’s circuit rather than the client’s. The client captures part of the benefit through the contract price, since competition among contractors presses their margins toward the point where the client’s saving and the contractor’s profit are both carved from the workers’ conditions. Analysts who measure only the client firm will announce that exploitation within it has fallen. The workers cleaning the same building at night, on worse terms, for a different logo, are the refutation.

Subcontracting chains extend the effect and dilute responsibility along with it. Each tier takes a margin, each tier pressures the next, and the final tier, often a small firm or a gangmaster-like intermediary employing the most vulnerable workers, operates with margins so thin that legality itself becomes a cost it struggles to afford. The Marxian analysis locates the surplus extraction at every tier where wage labour is employed, and locates the chain’s driving force at the top, where the lead firm’s buying terms set the prices within which every lower tier must find its margin. This is why the theory directs attention upward, to the firm whose name is on the product, rather than distributing blame evenly across the tiers in proportion to their visibility.

Franchising applies a related division to ownership itself. The franchisee is a small capitalist who employs workers, extracts surplus value in the local circuit, and then surrenders a substantial share upward as franchise fees and royalties, alongside compulsory purchases from the franchisor’s supply chain at prices the franchisor sets. The franchisor’s income is correspondingly a compound: rent on the brand, commercial profit on supplies, and a contractual slice of the surplus produced by the franchisee’s staff. The worker behind the franchised counter stands at the end of two extractions, one by the employer who sets the rota and one by the brand that sets the terms under which the employer operates. An analysis confined to the local employer misses the second extraction entirely, and an analysis confined to the famous brand misses the first. The arrangement is best understood as the franchisor renting out a proven circuit of accumulation to smaller capitals, keeping the property-like incomes and transferring the risks of employment, rent, and local demand downward.

Remote and distributed work moves the boundary in a different direction, into the worker’s home. When an employer closes offices and directs staff to work from their own rooms, on their own connections, often on their own equipment, a portion of the constant capital the firm once supplied, premises, heating, furniture, connectivity, is quietly transferred to the workforce, while the product and its surplus value continue to flow to the firm. The value of labour power does not automatically adjust to cover the transferred costs, and the working day becomes harder to bound when the workplace is also the kitchen table, which creates new scope for surplus labour in the form of the message answered at night and the task finished after the nominal close. Set against this, the worker recovers commuting time that was always an unpaid annex of the working day, and may reclaim part of it as genuine free time. The Marxian reading holds both effects in view: the firm’s costs fall and its claim on the worker’s hours potentially lengthens, while a real deduction from the worker’s day, the journey, is abolished. Which effect dominates for a given workforce is an empirical question about hours actually worked and costs actually borne, and it cannot be settled by the location of the desk.

The strongest objection: the productive and unproductive distinction may be unworkable

Every framework used in this article leans at some point on the line between productive and unproductive labour, so the objection to that line must be heard at full strength, in the form its most serious critics would choose rather than in a weakened version arranged for easy rebuttal. The objection has four strands. Together they amount to the charge that the distinction is an arbitrary inheritance from an economy of factories and domestic servants, unworkable in an economy of integrated services, and retained by parts of the Marxist tradition out of loyalty rather than usefulness.

The first strand is arbitrariness at the boundary. A distinction that classifies the restaurant cook as productive and the household cook as unproductive, though the meals are identical, appears to classify bookkeeping arrangements rather than anything in the world. The same training course is productive when a commercial college sells it and unproductive when a public institution provides it. The same lorry journey is production when a haulage firm sells transport and something murkier when a manufacturer’s own fleet moves its own goods. A category that flips with the employer’s legal form, the critic argues, is describing contracts, not labour, and an economics built on it mistakes the paperwork for the process.

The second strand is unworkability inside integrated work. The nineteenth-century firm separated its functions visibly: a counting house here, a workshop there. The modern professional employee moves between production, selling, administration, and compliance within a single week, sometimes within a single meeting. Classification would need time-use data at a granularity no firm records and no researcher can obtain, and different plausible allocations of the same person’s hours produce wildly different estimates of a firm’s productive wage bill. A measure whose inputs cannot be observed is not a measure. It is a ritual performed with numbers.

The third strand is empirical consequence. In the national-level literature on measuring the rate of exploitation, the productive and unproductive split is the single largest source of disagreement between published estimates, larger than any dispute about the underlying data. Researchers who draw the line differently reach different levels and sometimes different trends for the same economy in the same years. A distinction that drives results more powerfully than evidence does, the critic concludes, has ceased to be a discovery about the economy and become a dial on the analyst’s instrument, set by theoretical taste. Many Marxist economists have accordingly abandoned the distinction in their empirical work, treating all wage labour as exploited in a broader sense and measuring labour’s share and its movements instead, and their defection cannot be dismissed as ignorance of the texts.

The fourth strand is political and moral discomfort. However carefully the technical meaning is explained, the vocabulary lands on real workers as a ranking. Telling retail staff, carers, teachers in public institutions, and civil servants that their labour is unproductive in the master’s own terminology predictably alienates the very people a Marxist politics hopes to organise, and predictably delights opponents who quote the classification without its definition. A term that must be apologised for every time it is used, the critic says, costs more than it earns.

The reply begins by conceding the third strand almost entire. The classification decision does drive empirical results more than the data do, and this article has relied on that concession repeatedly: it is the deep reason behind the accounts warning. Any writer who reports a rate of exploitation for a firm or an economy without foregrounding the classification choices behind it is misleading the reader, and the objection performs a service by making that inescapable. The reply also concedes part of the second strand. Task-level classification of integrated white-collar work involves judgement calls that honest analysts will make differently, and the resulting estimates carry wide margins that should be reported as ranges or not reported at all.

What survives the concessions is the distinction’s analytical use, which does not depend on computing exact rates. Dropping the line between labour that produces surplus value and labour paid out of it dissolves several explanations that this article has needed. Without it, the retailer’s profit has no source but a markup, and the analysis of supply chains loses its account of how value produced in one place sustains incomes in another. Without it, the growth of employment in circulation, finance, and administration appears as a simple expansion of production, and the question of how a society supports a rising share of its workforce from a surplus produced by a falling share cannot even be posed. Without it, the difference between a firm investing in another production line and a firm investing in a larger sales apparatus vanishes, though the two decisions have opposite relations to the pool of surplus value. The distinction is retained in serious empirical work, by scholars fully aware of its costs, because it identifies a real structural feature: some activities add to the fund from which all incomes are drawn, and others draw from the fund, and an economy’s possibilities depend on the proportion between them. The verdict of this article is therefore conditional retention. Use the distinction to structure analysis, to locate firms and departments within the circuit, and to explain the movement of profit between sectors. Refuse to use it as the hidden machinery inside a falsely precise percentage. Judged by that standard the distinction is a good map and a bad oracle, and most of the damage attributed to it has been done by writers who demanded oracular work from it.

The competing account: profit as the reward for risk, coordination, and capital services

A fair application of the framework requires the rival explanation of the firm’s surplus to stand in its own best form, since a Marxian analysis that has never faced the strongest alternative is merely a vocabulary in search of converts. The mainstream account of profit, in its serious rather than its journalistic version, runs as follows. Production requires that resources be committed before revenue arrives, under uncertainty about whether anyone will buy the result. Somebody must advance the funds, bear the possibility of loss, coordinate specialised inputs that no market assembles unaided, and wait. Profit is the residual that rewards those functions: a return on capital comparable to interest, a wage for the coordinating and entrepreneurial labour of the firm’s organisers, and a premium for bearing risks that employees, paid by contract regardless of sales, have declined to bear. On this account the residual is not extracted from workers. It is what remains after all contributors, workers included, have received the contractually agreed price of their contribution, and its size measures how well the firm guessed what customers would value.

The account has genuine explanatory successes that a Marxian treatment should acknowledge without flinching. It explains why profits differ persistently between firms in ways that track product quality, timing, and organisational skill rather than workforce size. It explains why most new firms fail and why the prospect of the residual draws effort and capital into ventures whose workers, prudently, would never fund them from wages. It handles the software case with particular ease, since a product that customers value highly and that costs little to reproduce will obviously generate a large residual, with no need to invoke rent or transfers. And it rests on an observation that is simply true: the wage is paid whether or not the product sells, so the risk of the interval between production and sale is genuinely carried by whoever advanced the money.

The Marxian reply does not deny the interval, the uncertainty, or the coordination. It denies that bearing them explains the source of the residual. Risk is borne, in various forms, by workers too, whose livelihoods depend on the firm’s survival and who cannot diversify their single job across a portfolio of employers the way an investor diversifies holdings. More fundamentally, the reply distinguishes the question of who receives the residual and why institutions assign it to them, which the mainstream account answers well, from the question of where the value embodied in the residual originates, which it does not address. A firm’s revenue exceeds its costs, when it does, because the commodities sold embody more value than the inputs consumed, and the only input whose use produces more value than it costs is labour power. The mainstream account describes the distribution of the excess among claimants and the competition that disciplines it. The Marxian account claims to identify its production. The two are not equally deep answers to one question; they are answers to different questions that share a vocabulary, and much sterile controversy comes from treating them as rivals for the same ground. Where they do collide is over the wage: the mainstream treats the wage as the full price of labour’s contribution, while the Marxian insists the wage is the price of labour power, a different and smaller thing than the labour actually performed. That single distinction, labour against labour power, is the fork at which the two accounts separate, and no evidence about risk, however accurate, can adjudicate it, because both sides accept the same facts about who waits and who might lose.

A worked illustration: one imaginary firm, taken apart slowly

Abstraction has carried the argument far enough that a concrete pass through a single invented firm will fix the method in memory. The firm is imaginary, its numbers are round and illustrative, and nothing in what follows describes any real company. Call it a maker of kitchen equipment with a single plant, a small chain of branded showrooms, a service arm, and an arrangement under which most of its output is also sold through independent retailers. In a year it pays out, in round terms, sixty units of money in production wages at the plant, twenty in showroom and sales salaries, ten in administration and finance staff, and ten in managerial salaries that blend coordination with control. It spends heavily on materials and on machinery, which wears and is replaced. It pays interest to its bank, rent on the showrooms, and a royalty to the holder of a design patent it licenses. It ends the year with an accounting profit that commentators compare, in the usual way, with the total wage bill of one hundred, producing a reassuringly modest ratio.

Take the firm apart in the order the theory prescribes. The sixty units of production wages are variable capital, the only expenditure in the list that buys labour power engaged in producing value. The value those workers produce across the year is larger than sixty by exactly the surplus value they create, which the accounts nowhere state. The twenty units of showroom salaries are a cost of circulation, paid out of surplus value, though the showroom staff work surplus hours beyond the equivalent of their pay and thereby reduce the deduction. The ten units of administration divide between functions any production system needs and functions specific to selling, tax, and compliance. The ten units of management divide, in proportions no outsider can observe, between coordination that belongs to the collective labourer and control that belongs to the extraction process, with any equity element belonging to appropriation outright. Interest, showroom rent, and the design royalty are three further slices of the same surplus, claimed by a bank, by property owners, and by the holder of an intellectual right. The reported profit is the slice that remains with the firm after every other claimant has been fed.

Now observe what the popular ratio did. It divided the last slice by the sum of every salary in the building, plant and showroom and office together, and called the result the rate at which the firm exploits its workers. The numerator omitted most of the surplus, having surrendered interest, rent, royalty, and the circulation wages before the profit line. The denominator included tens of units of expenditure that are not variable capital at all. The two errors do not cancel; they compound in unknowable proportions, since a differently financed twin of this firm, owning its showrooms and borrowing nothing, would report a far larger profit on identical production and be ranked far more exploitative by the ratio while its plant workers work identical days. A method that ranks twins differently because of their landlords and lenders is measuring the property and credit system. It is not measuring the working day. Recording a structured set of notes on a real firm in this disaggregated form, rather than as a single ratio, is the practical habit this article recommends, and a study tool such as VaultBook’s Marxism study notes serves that habit well, since the classification of each cost line can be kept, revised, and argued with instead of being collapsed into one indefensible percentage.

Services, care, and the state contractor: where the distinction earns its keep

Service businesses test the framework differently from factories and shops, because the product is an activity consumed as it is performed. A private clinic, a commercial training provider, a chain of nurseries, a contract cleaning company, and a firm running outsourced public services all sell labour’s direct exercise rather than objects that outlast it. Marx’s treatment accommodates this without strain: a service is a use value whose production and consumption coincide, and labour employed by capital to produce services for sale is productive labour when the sale valorises the capital advanced. The teacher in the fee-charging school was his example, and it generalises cleanly to the instructor in a commercial training firm, the clinician in a for-profit clinic whose treatments are sold, and the nursery worker in a chain whose places are sold to parents by the month. In each case the working period divides into necessary and surplus portions, and the firm’s margin is fed by the division, alongside whatever rents its locations, licences, and accreditations command.

The same work performed under public provision classifies differently, and the difference exposes what the category is really about. A nurse in a public hospital and a nurse in a private one perform indistinguishable labour for indistinguishable patients. The first is paid from taxation, which is a deduction from the total surplus value and wages of the economy, and produces no surplus value. The second is employed by a capital and does. A government that contracts the hospital’s cleaning to a private firm has not changed the cleaning; it has changed the social form, moving the cleaners from unproductive public employment into productive employment by a contractor, where their labour now yields a margin for the contractor’s owners as well as clean wards. Whether that change is desirable is a political question this article does not settle. That the change alters the classification, and alters where the surplus labour of the cleaners becomes visible as someone’s profit, is an analytical fact, and it explains why privatisation debates so often talk past each other: one side describes the work, which is unchanged, while the other describes the circuit, which is transformed.

Care work presses a further question that the framework must answer carefully. Much care is performed outside the wage relation entirely, in households, unpaid, and it reproduces labour power daily and generationally. That labour produces no value in Marx’s technical sense, because it produces no commodities for sale, yet without it the value of labour power could not be sustained at its given level and the whole structure of necessary labour would shift. A firm-level article is not the place to develop the theory of social reproduction, but the boundary matters for analysing care businesses: when a firm sells care that families once provided unpaid, labour that stood outside the circuit moves inside it, becomes productive in the capitalist sense, and begins to generate surplus value, while the total care performed in society may not have changed at all. Growth in the measured productive sector can therefore reflect the migration of activity across the market boundary rather than any increase in the activity itself, a warning that applies equally to commercial childcare, meal delivery replacing home cooking, and platforms selling domestic services by the hour.

The small firm and the self-employed: extraction without an employer

At the lower edge of the firm size scale the analysis meets businesses where owner and worker are the same person, or where a handful of employees work beside an owner who also labours. The framework handles the sole trader by a clean division: a self-employed person who sells a product or a service appropriates the whole product of the labour, necessary and surplus portions together, and is exploited by no employer, however long the hours worked and however meagre the return. Self-exploitation is a real discipline, imposed by prices, rents, platform fees, and competition from larger capitals, but it is not the extraction of surplus value by another party, and precision here protects the concept from becoming a synonym for finding work hard. The moment the same trader hires an assistant, a small circuit exists: the assistant’s labour power is bought, a surplus portion appears, and the owner appropriates it, perhaps while also performing surplus labour alongside the assistant in the same workshop.

Small firms matter to the larger analysis for two structural reasons. First, they are the seedbed and the graveyard of capitals, the zone where the pressures described throughout this article operate with the least cushioning: the small manufacturer squeezed between a dominant retailer’s prices and a landlord’s rent experiences the distribution of surplus value as a weekly emergency rather than a theory. Second, large firms deliberately maintain a penumbra of nominally independent small suppliers, contractors, and franchisees whose workers’ surplus value the large firm partly captures through prices and fees while bearing none of the obligations of direct employment. The boundary of the firm, celebrated in management literature as a strategic choice about core competence, appears in this analysis as a choice about where extraction is located and how much of its yield can be captured without owning the workplace in which it occurs. Asking of any large company not only how it treats its employees but how its prices, fees, and contract terms treat the employees of its dependents is one of the most productive habits the framework teaches, and it follows directly from the supply chain and franchise discussions above.

What can legitimately be inferred about a real firm

The accounts warning is a prohibition on one kind of claim, not a prohibition on knowledge, and this section gathers the positive programme that the preceding analysis licenses. An analyst who respects the warning can still establish a great deal about a firm, in descending order of confidence. At the highest confidence, the analyst can describe the business model’s structure: which activities the firm performs that produce value, which realise it, which collect shares of it as interest, rent, royalty, or commercial profit, and how the proportions among those activities have shifted as the firm has outsourced, franchised, automated, or moved into finance. Business models are public facts, stated in annual reports and visible in operations, and classifying them is the framework’s home ground.

At the next level of confidence, the analyst can identify the mechanisms bearing on the rate of surplus value for each workforce the firm commands directly or indirectly. Lengthening or fragmenting hours points one way; intensification through monitoring, targets, and algorithmic pacing points the same way; genuine productivity growth in wage goods points the same way by a different route. Falling real wages against rising productivity in the firm’s own operations point to a rising rate with fair confidence, since the division of the working day is being squeezed from both ends. Rising dependence on fees charged to nominally independent workers points to extraction relocating rather than shrinking. None of these inferences requires the accounts to disclose value magnitudes. They require evidence about hours, intensity, staffing, and terms, much of it obtainable from workers’ own testimony, tribunal records, procurement documents, and the firm’s operational literature.

At the third level, the analyst can reason about the firm’s position in the distribution of the social pool. A firm whose profits persistently exceed what its own productive workforce could plausibly generate is capturing value produced elsewhere, and the channels of capture, buying power, intellectual property, property, lending, or brand, can usually be named. A firm whose profits are thin despite a large and intensively worked productive workforce is surrendering its surplus to claimants along its circuit, and those claimants can usually be named too. These are judgements about flows and positions, not measurements of rates, and they are the level at which the framework does its most valuable public work, because they redirect scrutiny from the single ratio in a headline to the structure that produced the headline. What the analyst must refuse, at every level, is the conversion of these structural findings into a spurious percentage. The refusal is not modesty about this article’s method. It is the method, since the credibility of everything claimed here depends on declining the one claim the evidence cannot support.

Hours, intensity, and the monitored workplace: the levers as they now operate

The oldest lever on surplus value is the length of the working period, and in economies where the standard week is legally bounded the lever has changed its grip rather than disappearing. It operates through unpaid extensions at the edges of the paid day: the shift that begins with preparation the roster does not count, the logout procedure completed after clocking out, the messages answered from home, the availability demanded of workers paid only for tasks completed. Each fragment is small enough to escape individual notice and large enough, multiplied across a workforce and a year, to move the division of the total working time materially. Marx treated the working day as a contested magnitude whose boundaries are set by struggle and law rather than by physiology alone, and the fragmentation of the modern working period has multiplied the number of boundaries to contest. A workforce that wins a shorter roster while surrendering its evenings to availability may have traded a visible concession for an invisible extension, and only an analysis that follows the labour rather than the contract can detect the exchange.

Intensity is the second lever, and the one whose importance has grown most as hours have been regulated. A worker who serves more customers per hour, picks more items per shift, drives more deliveries per evening, or resolves more cases per day expends more labour in the same clock time, and the surplus portion grows without any change in the roster. Digital monitoring has given this lever a precision earlier managers could not approach. Where the foreman’s eye sampled a floor intermittently, the dashboard samples continuously: pick rates, call durations, idle intervals, route deviations, keystroke rhythms. The measurement itself becomes a pacing instrument, because a worker who can see a personal rate displayed against a target adjusts effort before any supervisor intervenes. From the standpoint of value theory, intensification occupies an instructive middle ground between the absolute and relative forms: like lengthening, it extracts more labour from the same worker in the same post, and like a productivity rise, it cheapens each unit of output and can spread through an industry as the standard pace, resetting what counts as a normal hour for everyone. The worker’s experience registers the distinction poorly, since an intensified hour and a lengthened day both arrive as exhaustion, but the analyst needs it, because the two levers meet different legal and physical limits and provoke different forms of resistance.

The third lever is the cheapest and the quietest: lowering the value of labour power itself by cheapening the goods and services workers consume. No single firm controls this lever, since it operates through productivity across the whole wage goods sector, yet firms benefit from it collectively and some large employers reach toward it directly, through staff discount schemes, subsidised canteens, tied housing in earlier eras, and benefits structured to substitute for wage increases. When the cost of reproducing the workforce falls, the necessary portion of every working day in the economy shrinks, and surplus labour expands without a single roster changing. This is the mechanism that makes the rate of exploitation a social magnitude before it is a firm-level one, and it is the deepest reason firm-by-firm league tables mislead: a company whose workers’ rate has risen may have done nothing at all, merely employing labour in an economy where food, clothing, and communication grew cheaper around it.

The firm as lender and landlord: income that arrives as a share

Large firms increasingly earn income from positions that produce no value in their own right, and the analysis of a modern company is incomplete until those positions are separated from its production. A manufacturer that finances its customers’ purchases collects interest. A retailer that issues credit collects interest and fees. A franchisor collects what functions as rent on a brand. A group that owns its premises through a property subsidiary charges its operating companies rent, moving surplus between its own pockets in ways that flatter or depress each unit’s reported performance according to internal decisions. An employer that runs a pension fund, a leasing arm, or an insurance scheme for its supply chain stands in several of these relations at once. Each income stream is a claim on surplus value produced somewhere in the economy, and the firm’s consolidated result blends its earnings as a producer with its takings as a creditor, a landlord, and a licensor.

The separation matters because the streams behave differently and fail differently. Profit on production depends on the firm’s workers and markets. Interest depends on the solvency of borrowers across the economy. Rent on intellectual property depends on legal enforcement and on the scarcity the law maintains. A firm can therefore grow more prosperous while its production stagnates, by expanding the share-earning side of its structure, and commentary that reads its rising profit as evidence of rising exploitation of its own staff mistakes the direction of travel. The reverse also occurs: a firm may squeeze its productive workforce harder each year while its reported result deteriorates under interest payments on debt taken on for acquisitions, so that the workers’ rising surplus labour services the firm’s creditors rather than its owners’ dividends. Both patterns are commonplaces of corporate life, and both are invisible until the surplus is followed as a flow with multiple claimants rather than admired as a single figure at the foot of an income statement.

Reading a set of accounts against the grain

Since the accounts cannot be avoided, only demoted, this section sets out the order in which a careful reader can interrogate them for the limited purposes they can serve. Begin with the structure of the business rather than its totals: the segment reporting, where groups disclose the activities they are divided into, usually separates production from services, retail, finance, and property more honestly than the consolidated lines do, and the analyst’s first task is to assign each segment a place on the circuit described at the start of this article. Follow with the employment disclosures, noting not the total wage bill but its distribution across segments where given, and treating any single consolidated staff cost figure as unusable for value analysis. Note the interest paid and received, the rental obligations, and the royalty flows in both directions, since each measures a channel through which surplus value enters or leaves the firm’s own production. Note the jurisdictions named in the tax disclosures, and hold them beside any public knowledge of where the firm’s plants, warehouses, and workforces actually are, because the distance between the two lists is the geographical face of the accounts warning. Note depreciation policies and any changes to them, and discount profit movements that coincide with a change of schedule.

What emerges from such a reading is not a rate but a map: a picture of a firm as a bundle of circuits and claims, with the productive core identified, the realisation apparatus measured as a deduction, and the share-earning positions flagged. Maps of this kind can be compared across firms and across years with real profit to the reader, because changes in the map correspond to changes in the world: a rising share of income from financing, a shrinking productive segment inside a growing group, a lengthening list of low-tax booking locations beside an unchanged list of production sites. Each of those is a finding. None requires pretending that the map’s legend contains a percentage sign. The discipline this article has urged from its opening section can now be stated as a working rule for anyone who writes about companies: describe the circuit, name the claimants, identify the levers on the working day, and let the rate go uncomputed, because the evidence for it is not in the documents and the documents were never designed to hold it.

What this analysis sees that a standard reading of a company misses

The case for doing all this work, when the reward is a map rather than a number, rests on what the map shows that conventional company analysis leaves out. A standard reading of a firm asks how its margin is defended: pricing power, cost discipline, brand strength, switching costs, scale. The Marxian reading asks a prior question that the standard reading cannot formulate, which is where the margin’s substance originates and whose unpaid time ultimately composes it. That question changes the significance of facts the standard reading also possesses. A cost programme that a business page reports as efficiency appears here as a redistribution of the working day. A supply chain celebrated for asset-light elegance appears here as a device for locating production where the value of labour power is lowest while locating capture where property is strongest. An asset-light firm, in particular, deserves the second look this framework gives it, because lightness of assets usually means heaviness of claims: the firm has shed the constant capital and the direct employment while retaining, through contracts, brands, and platforms, a senior claim on the surplus that other firms’ workers produce.

The analysis also restores time to the centre of the description of a firm. Conventional accounts of a company are written in money and ratios, and the working day, the substance being divided in every workplace the firm touches, vanishes from them entirely. Restoring it explains phenomena that money language handles awkwardly: why firms resist reductions in hours with an intensity out of proportion to the wage saving at stake, why monitoring technology attracts investment beyond its measurable effect on output, why the scheduling of fragments of labour has become a discipline of its own. Each of these is a contest over the boundary between necessary and surplus time, conducted with software and rosters rather than with the factory whistle. A reader equipped with the division of the day can see those contests in a quarterly report’s bland references to labour flexibility, which is perhaps the most practical literacy this article can confer.

Where the framework strains, and what the strain teaches

Honesty about limits has been this article’s refrain, and the remaining strains should be collected in one place rather than left scattered. The first strain is intangible production. Where a firm’s chief asset is a network, a dataset, a standard, or a reputation, the share of its income attributable to the current labour of its employees shrinks relative to the share attributable to position, and the theory’s rent categories absorb so much of the result that the production analysis, though still valid for the workforce, explains a minority of the income. Critics inside and outside the tradition have proposed treating user activity itself, the content, attention, and data that platform users supply without wages, as a form of labour producing value. This article records the proposal without adopting it: the activity is real, the appropriation is real, and the analogy to the working day illuminates the firm’s dependence on unpaid contribution, but calling every monetised activity labour dissolves the precision that makes the category worth having, and the stricter treatment of such income as rent on a monopolised space fits the evidence of how the income behaves. Readers should know the dispute exists and that it marks a genuine frontier rather than a settled province.

The second strain is the household and the boundary of the market, raised earlier in connection with care. As services migrate between unpaid domestic provision, public provision, and commercial provision, measured production and measured exploitation move for reasons that are partly bookkeeping about social form. The theory can describe each migration exactly, and indeed the description is one of its strengths, but economy-wide comparisons across long periods inherit the instability, and firm-level snapshots inherit it too whenever a company is praised for growth that consists of absorbing what households or the state once did directly.

The third strain is the global one. Where production, realisation, and booking occur in different jurisdictions under different wage levels and different legal forms of the working day, the unit of analysis quietly dissolves. The firm is a legal fact; the circuit is an international fact; the accounts are a negotiated fact. Analysis at the level of the firm therefore has to proceed, as this article has, by following the circuit through the firm rather than by treating the firm as a sealed container of value. That is a modification of the classical presentation, in which the individual capital could be treated as a self-contained example, and it should be owned as a modification. The categories survive the journey. The unit they were first illustrated on does not, and pretending otherwise produces the confident national or corporate league tables that this article has spent its length refusing to build.

Wage forms in the modern firm: salary, piece rates, bonuses, and equity

The form in which labour power is paid shapes how visible the division of the working day remains, and modern firms deploy a repertoire of forms whose variety conceals a common function. The monthly salary paid to professional staff presents the wage as a price for a role rather than for hours, and thereby dissolves the boundary of the working day altogether: the salaried engineer whose contract names no hours can be assigned the evening deployment and the weekend incident without any transaction occurring in which extra labour is visibly exchanged for nothing. Marx treated the salary of commercial and supervisory staff as a form especially prone to this illusion, because a payment attached to a person and a status looks nothing like the purchase of a divisible quantity of labour power, though the quantity is what the employer is managing. The growth of salaried employment is therefore, among other things, a growth in the opacity of the surplus portion, and one reason white-collar workforces often discover their own exploitation later and describe it less precisely than workforces paid by the hour.

Payment by results moves in the opposite direction and conceals by compression rather than by dissolution. The piece rate, the per-delivery fee, the commission on sales, and the bonus tied to output all present the wage as a share of product, so that the worker appears to be paid for work done rather than for time surrendered. The appearance is powerful, because it makes an increase in effort look like the worker’s own entrepreneurial choice, while the rate itself is set and reset by the employer in light of the effort it elicits: when workers earn too easily at a given piece rate, the rate is revised, a sequence documented in industry after industry since the nineteenth century. Bonuses and equity add a further turn. A bonus paid from profits invites the worker to identify with the residual, and an allocation of shares converts a slice of the surplus into part of the worker’s own income, which can align a favoured layer of employees with the extraction performed on the rest and on themselves. None of these forms alters the underlying magnitudes: in every case labour power is bought, labour is performed beyond its value, and the form determines mainly how much of that structure the worker can see from inside it. Analysts should therefore treat a firm’s compensation architecture as evidence about concealment and alignment, while looking for the rate itself in the division of time and the value of the wage, where it has always lived.

The warehouse and the road: logistics as production continued by other means

Between the factory and the shop stands the sector that modern retail has made enormous and that casual analysis usually forgets: warehousing, freight, and delivery. Marx’s classification of transport is unambiguous and repays attention, because it contradicts the instinct that only making things counts. Transport is a branch of production, he argued, because the use value of a commodity is realised only where it can be consumed, and changing a thing’s location is as much a material transformation of its usefulness as changing its shape. The lorry driver moving components to an assembly plant and the driver moving finished goods to a distribution centre perform productive labour. Storage is productive insofar as it preserves use values across the interval between production and sale, and unproductive insofar as stocks pile up because circulation has stalled. The warehouse worker who receives, sorts, picks, and dispatches is therefore not a retail worker who happens to stand in a larger building. Much of that labour continues production, and the surplus value it yields belongs to the same account as the factory’s.

The modern fulfilment centre makes an instructive specimen because it fuses this productive core with the most developed apparatus of intensification described earlier in this article. Labour there is measured in units per hour by systems that know the location of every item and the position of every worker, and the working day is organised around rates that adjust to demand by the hour. The classification and the experience illuminate each other: the theory says this labour produces value and surplus value, and the organisation of the workplace, with its relentless attention to the division of the hour, behaves exactly as a workplace behaves when the surplus portion of each hour is the quantity being managed. Compare the shop floor of the retail store the same company owns, where staffing follows the different logic of minimising a deduction. The two workplaces feel different to the people in them, in ways that are hard to articulate, and the difference in their position on the circuit, producing value in one case and realising it in the other, is a large part of what they are feeling.

Logistics also concentrates the geography lesson of the supply chain section into a single industry. Ports, freight corridors, and distribution parks are sited where land is cheap, labour is available, and planning law is permissive, which rarely coincides with where the goods are consumed or where the owning firm’s profit is booked. The value produced by dockers, drivers, and pickers is therefore generated in one set of places, realised in another, and reported in a third, and the logistics firm that employs them may itself be a contractor two tiers below the brand whose goods fill its shelves. An analyst following a single parcel backward from a doorstep can traverse four or five distinct circuits of capital, each extracting or claiming a share, before reaching the factory where the contents were made. That traversal, unglamorous as it sounds, is what applying this framework to a modern firm actually consists of, and no shortcut through the accounts of any one company on the route can substitute for it.

For students and exam candidates

Use modern firms in answers as illustrations of mechanisms, never as evidence for magnitudes. An examiner rewards the candidate who shows a software company’s working day dividing into necessary and surplus portions and who then states, precisely, why the company’s reported margin cannot be converted into a rate of exploitation. The same examiner penalises the candidate who divides a famous firm’s profit by its wage bill, however topical the example feels. Choose firm types rather than named companies wherever possible: the manufacturer, the retailer, the platform. Types cannot be fact-checked into embarrassment, they keep the argument at the level of structure where marks are earned, and they let you deploy the accounts warning as a sign of sophistication rather than as an evasion. Where a question invites a contemporary example, give the mechanism first, the example second, and the limitation third, in that order, every time. A short answer that names the four reasons accounts fail will outscore a long answer that merely repeats the definition of surplus value with a famous company’s name attached.

For teachers and seminar leaders

The productive and unproductive distinction is the most teachable element in this material because it reliably produces a productive quarrel. Set the class the paired cases: the cook in the restaurant and the cook in the household, the nurse in the private clinic and the nurse in the public hospital, the driver who owns the car. Require students to classify each case and to name the criterion they used, then require them to defend the classification against the arbitrariness objection before revealing how much of the later argument depends on it. A second exercise uses the worked illustration in this article: give the seminar the imaginary firm’s cost lines, ask each group to compute the popular ratio, and then ask what the ratio omitted and double-counted. Students who have performed the error themselves, and dismantled it themselves, retain the accounts warning far better than students who have merely been told it. Keep the political temperature governed by the technical framing: the question throughout is what the categories classify, and the class that argues about classification is learning the material.

For researchers, historians, and journalists

For researchers, the open problem this article leaves is empirical design: a defensible firm-level study would need task-level time data, segment accounts restated on Marxian categories, and the interest, rent, and royalty flows traced as shares, and no standard dataset supplies any of the three. Studies that proceed anyway should publish their classification rules alongside their results, since the rules are the results, and historians can contribute by documenting how firms themselves have understood the division of the working day in their own internal records, where the language of labour cost per unit often tracks the theory more closely than the published accounts do. For journalists and analysts, the practical rules are stricter because the audience is larger. Never convert a margin into an exploitation rate. Never describe a company’s profit as the amount its workers were underpaid, since the profit is a share of a pool and the workers’ surplus is a different magnitude that the accounts do not state. Do verify the structure before writing: where the firm produces, where it books, who supplies its labour directly and through contractors, which of its incomes are rents and interest. A report that names the mechanism, the claimants, and the limits of the evidence will be less dramatic than a fabricated percentage and considerably harder to refute, which in this contested territory is the form authority takes.

What the analysis establishes, and what remains open

The application of the surplus value framework to firms unlike any Marx inspected yields a mixed and therefore credible result. The mechanism survives intact: wherever labour power is bought and set to work producing commodities or services for sale, the working period divides into necessary and surplus portions, and the division can be traced through software teams, clinics, nurseries, kitchens, and warehouses as reliably as through mills. The distribution analysis survives and strengthens: the modern firm’s profit is visibly a share in a wider pool, claimed alongside banks, landlords, licensors, franchisors, and platforms, and the framework’s insistence on following the surplus as a flow explains the shape of supply chains, franchise systems, and asset-light business models better than any account that treats each firm’s profit as self-generated. The measurement ambition does not survive, and its failure is the article’s central finding: between the consolidated accounts and the Marxian categories stand the mixed wage line, the transferred shares, the chosen booking jurisdiction, and the depreciation convention, and no ingenuity applied to published documents can remove them. What remains open is equally definite. The classification of platform labour turns on facts about control that vary by design and by jurisdiction. The treatment of income from intangibles presses the rent categories harder than they were built to bear. The classification of integrated white-collar work resists the granularity the theory ideally wants. An analysis that ends with its open questions named, rather than dissolved, has applied the framework the way the framework asks to be applied: as an account of a real structure, confident where the structure is clear and silent where the evidence runs out.

Seasonal, casual, and agency labour: the working day bought in slices

A further feature of the modern firm deserves its own examination, because it alters the form of the wage relation without altering its content: the purchase of labour power in slices too thin to support the worker who sells them. The seasonal worker hired for a harvest or a holiday peak, the agency worker supplied to a warehouse for a single project’s duration, the casual staffer engaged for a venue’s busy evenings, and the zero-hours employee who must remain available without any guaranteed purchase all sell labour power under contracts that commit the buyer to almost nothing beyond the hours actually used. Marx analysed piece wages and day wages as forms; the thin contract is a further form, whose distinctive work is to transfer the risk of fluctuating demand from the firm to the worker. Demand for the firm’s product rises and falls, and under a full employment contract the firm bears the cost of the troughs in idle paid time. Under the thin contract the troughs become the worker’s unpaid waiting, and the firm’s variable capital is advanced only for hours that production or realisation actually fills.

The effect on the extraction of surplus value runs through several channels at once. Hours bought only when needed are bought at moments of peak intensity, since the casual worker is summoned precisely when the pace is highest and dismissed when it slackens, so the surplus portion of the hours actually purchased tends to be large. The value of the worker’s labour power, meanwhile, is still set by the cost of maintaining the worker across the whole week and year, while the wage arrives only for fragments of it, and the gap between the cost of staying alive and available and the income the fragments yield is borne by the worker, by other household earners, or by public support. A firm can therefore present a wage rate that looks respectable per hour while the arrangement as a whole pays the workforce, across the year, less than the value of the labour power it consumes, the difference being made good outside the firm’s accounts entirely. Agency intermediation adds one more claimant to the circuit: the agency’s margin is a deduction from the flow between the firm and the worker, earned for performing the matching and the risk-shifting, and it is paid, ultimately, out of the same surplus and the same wage fund that the two principals divide.

None of this makes casual workers a separate class or their labour a different kind. It makes their contract a technology for dividing the working day more finely than the standard contract allows, and for locating the costs of capitalism’s fluctuations on the party least able to refuse them. The analyst examining a firm with a large casual or agency workforce should read the low and stable labour costs in its accounts as a sign that those costs have been displaced, and should ask where they landed: on households absorbing irregular incomes, on workers holding several thin contracts at once, or on the public services that support people whose labour power is consumed in slices too small to live on. That question, like every question this article has posed, is answered by following the labour rather than the ledger.

Forms Marx already faced: the joint-stock company, the manager, and the cooperative factory

It is sometimes said that the framework belongs to an age of owner-managed mills and cannot speak to corporations whose owners are pension funds and whose direction is salaried. The objection underestimates its author. Volume Three of Capital contains an extended treatment of the joint-stock company, which Marx recognised as a transformation in the form of capital itself: ownership disperses among shareholders who supply money and nothing else, direction passes to employed managers, and profit splits visibly into interest-like dividends and the earnings of enterprise. He described the result as the abolition of capital as private property within the confines of the capitalist mode of production itself, a formulation that startles readers who expect him to have missed the corporation entirely. For the analysis of firms the consequences are direct. The separation of ownership from direction explains why the extraction of surplus value proceeds without any owner on the premises and without personal animus anywhere in the building: the manager who intensifies the working day is performing a function that the structure assigns, and would be replaced for declining to perform it. Exploitation in the modern corporation is not a residue of personal greed surviving into an impersonal age. It is what the impersonal structure is for, and the joint-stock form, by stripping away the owner’s person, makes that clearer than the family firm ever did.

The cooperative factory receives an equally striking treatment in the same volume, where Marx discusses worker-owned enterprises that had arisen within the old form. His judgement was double. Such factories demonstrate that the capitalist as a functionary of production can be dispensed with, since the workers, associated and directing themselves, run the process without him. At the same time, the workers become their own collective capitalist in relation to their own labour, employing their means of production to valorise their work under market conditions that still impose the old disciplines, so the antagonism is reproduced in a new seat rather than abolished outright. The modern employee-owned firm and the platform cooperative live inside exactly this double judgement. Their members may appropriate the whole product of their labour and divide it by rules they set, which removes the extraction of surplus value by an external owner. Their members may also lengthen their own working days and intensify their own labour to meet prices set by competitors who do extract surplus value in the ordinary way, which is why the theory insists that the market’s compulsions outlast any single firm’s constitution. Both observations are needed, and the framework’s ability to hold them together, praising the cooperative as a proof and analysing it as a prisoner of its environment in consecutive sentences, is a fair sample of the honesty this article has tried to practise throughout.

Frequently Asked Questions

Q: How does surplus value work in a software company?

Engineers, designers, and testers employed by the company sell their labour power for a wage and work a period that divides into a necessary portion, reproducing the value of that wage, and a surplus portion, whose product the company appropriates. The product being intangible changes nothing about the relation, since value follows the social form rather than the material. What does complicate the case is reproduction: code is copied at almost no cost, so much of the revenue behaves like rent on a legally fenced asset rather than payment for fresh production, and a careful analysis separates the productive labour of building and maintaining the service from the rent-like income its monopoly position commands.

Q: Can you calculate a company’s rate of exploitation from its accounts?

No, not reliably, and the failure is structural rather than a matter of missing disclosures. The wage line mixes productive workers, whose pay is variable capital, with circulation and administrative staff whose pay is a deduction from surplus value. Reported profit is what remains after interest, rent, and royalties, which are themselves shares of surplus value, have been paid away. Profit is booked where the group chooses to recognise it for tax purposes rather than where labour was performed. Depreciation follows conventions that move the result without touching production. Each defect alone would distort the popular profit-over-wages ratio. Together they make it a measure of accounting and financing arrangements, not of the working day.

Q: Are gig workers exploited if they own their own car?

Ownership of the vehicle does not settle the question, because the criterion is the relation, not the asset list. Where a platform sets the price, allocates the work, disciplines through ratings, and prevents the driver from building an independent clientele, the arrangement is a wage relation in substance, the commission is the form of extraction, and the driver’s ownership of the car mainly transfers fuel, maintenance, and depreciation costs from the firm to the worker. Where the driver sets prices, chooses clients, and works through several channels, the better description is a small producer paying rent for access to a monopolised market. The first driver is exploited in the technical sense. The second is squeezed by rent and competition, which is a different diagnosis requiring a different remedy.

Q: What is productive and unproductive labour?

Productive labour, in the capitalist sense Marx uses, is labour exchanged against capital that produces surplus value for its purchaser. Unproductive labour is labour exchanged against revenue, or paid out of surplus value, that produces no surplus value however useful it may be. The distinction classifies the social relation, not the worth of the work: a teacher in a fee-charging school is productive for the owner, a teacher in a public school is not, and the teaching may be identical. Inside firms the line separates production from pure circulation, selling, and money handling, with transport and maintenance falling on the productive side. The classification is indispensable for tracing where surplus value originates and contestable at the boundary of integrated white-collar work, where one salary covers both kinds of task.

Q: Where is value created in a global supply chain?

Value is created at every point where productive labour is performed under capital: extraction, component manufacture, assembly, and the transport that connects them, in proportion to socially necessary labour expended. It is not created in proportion to where the final price is largest or where profit is booked. Design, branding, and retail capture value rather than creating the bulk of it, drawing shares of the pool produced upstream through intellectual property, buying power, and control of the customer relationship. The characteristic structure of a branded chain is therefore production concentrated where the value of labour power is low and appropriation concentrated where property rights are strong. Booking the profit in a low-tax jurisdiction relocates the record of the capture, not the creation.

Q: Do managers produce surplus value?

Partly, and the division runs through each manager’s working week rather than between managers and workers. Insofar as a manager performs the coordination that any large combined labour process requires, sequencing work, solving problems, and holding a complex operation together, that activity belongs to the collective labourer and shares in production. Insofar as a manager performs control, monitoring effort, enforcing pace, and disciplining a workforce whose surplus labour is the point of the exercise, the salary is a cost of extraction paid out of surplus value. Senior executives paid partly in shares and profit-linked bonuses stand further across the line, since that portion of their income is not a wage at all but a direct claim on the surplus the firm appropriates.

Q: Is a retail worker productive labour?

In the strict sense, no. Selling changes the ownership of commodities rather than their value, so shop staff perform circulation labour and the retailer’s profit is commercial profit, a share of surplus value produced by workers upstream in manufacturing, agriculture, and transport. The classification is not a judgement on the work, which is often harder and worse paid than the production work it depends on. Retail workers also perform surplus labour in the sense that their working day exceeds the portion their wage represents, and that excess reduces the employer’s cost of realising value even though it creates none. Warehouse and delivery staff are treated differently, because transport and storage continue production into circulation and count as productive on Marx’s analysis.

Q: How does outsourcing affect surplus value?

Outsourcing relocates extraction rather than reducing it. When a firm replaces its own cleaners, drivers, caterers, or production workers with a contractor’s staff, the labour continues, the surplus portion of the working day continues, and the surplus value is now produced inside the contractor’s circuit. The client firm captures part of the benefit through the contract price, since competition among contractors presses margins down toward the workers’ conditions. The client’s own accounts then show a smaller wage bill and a leaner headcount, which careless analysis misreads as diminished exploitation within the firm. Chains of subcontractors multiply the effect, with each tier taking a margin and the pressure accumulating on the final workforce, while the lead firm’s buying terms set the prices within which every tier must operate.

Q: Is a highly automated factory more or less exploitative?

Per remaining worker, typically more; per unit of capital, less profitable, and the confusion between those two statements generates most of the muddle around automation. Machinery raises the output each worker sets in motion and, as productivity cheapens the goods workers consume, shrinks the necessary portion of the working day, so the rate of surplus value for the technicians who remain can be very high. The same machinery swells constant capital relative to variable capital, which presses the rate of profit downward even while the rate of exploitation rises. A small headcount therefore proves nothing about gentleness. It may indicate intense surplus labour extracted from a few, thin returns on a large capital, and a displaced workforce whose availability now disciplines wages across the industry.

Q: How does a brand capture value made in another country?

Through property and position rather than through production. The brand’s owners hold the design, the trademark, and the customer relationship, and those assets command payments at every stage: licence fees inside the corporate group, prices dictated to component suppliers and assemblers whose alternative customers are few, and a retail margin set by the strength of the name rather than by the labour performed near the point of sale. The surplus value produced by assembly and component workers, whose wages reflect the low value of labour power where they work, is thereby drawn toward the jurisdiction and the company where the brand sits, and booked there as the brand’s profit. No secret transfer is required. Contracts, intellectual property law, and unequal bargaining power perform the capture in the open.

Q: Does a loss-making platform company still organise the extraction of surplus value?

Yes, because extraction, realisation, and reported profit are three different events and a platform can separate them for years. Drivers and couriers perform surplus labour on every task whenever the relation is a wage relation in substance, regardless of the firm’s bottom line. A platform subsidising prices to build dominance realises the value of that labour below its full amount, transferring part of the benefit to customers as cheap journeys and spending another part, alongside investors’ funds, on purchasing a market position. The accounts show a loss financed by capital raised against expected future rents and extraction. Reading the loss as proof that no exploitation occurred reverses the facts: the workers’ surplus labour is one of the assets being spent to build the monopoly.

Q: How does surplus value analysis treat a franchise business?

As two linked extractions rather than one. The franchisee is a small capitalist who buys labour power locally, directs the work, and appropriates the surplus value the staff produce. A substantial share of that surplus is then surrendered upward as franchise fees and royalties, which function as rent on the brand, and through compulsory supply purchases priced by the franchisor, which add a commercial profit on top. The franchisor’s income is correspondingly a property-like claim on value produced by workers it does not employ, in premises it does not run. The worker at the counter stands at the end of both extractions: the rota set by the local employer and the terms set by the distant brand between them determine the wage, the pace, and the staffing level of every shift.

Q: What should an analyst verify before describing a company margin as exploitation?

First, what the margin actually contains: whether the firm’s income comes mainly from production, from realisation, or from interest, rent, royalties, and fees that are shares of value produced elsewhere. Second, where the labour is performed and by whom, including contractors, franchisees, and supply chain workers whom the firm’s own headcount excludes. Third, the levers bearing on the working day in those workplaces: hours actually worked, intensity, monitoring, and the movement of real wages against productivity. Fourth, where profit is booked relative to where production sits, since a wide gap between the two lists is the geographical signature of transfer. An analyst who has verified the structure can describe mechanisms and claimants with confidence. An analyst who has verified only the margin can describe the margin, and should stop there.

Q: Why do intellectual property rents complicate a surplus value analysis of software firms?

Because the income they generate does not correspond to labour performed in the period being analysed. A patent, a copyright over code, or a protected standard allows its holder to charge for access to something whose reproduction costs almost nothing, so the revenue stream behaves like rent on a scarce asset whose scarcity the law maintains. In the firm’s accounts that rent arrives blended with the proceeds of the engineers’ productive labour, and no published line separates them. Treating all the revenue as the product of the current workforce overstates the value produced in the period, while treating none of it as connected to labour ignores the workforce that built and maintains the fenced asset. The honest treatment names both components, assigns subscription income a rent element of unknowable size, and refuses the false precision of a split the documents cannot support.

Q: How does remote work change where surplus value is produced?

The location of production moves into the worker’s home while the appropriation stays exactly where it was, with the firm. Several quieter transfers accompany the move. Premises, heating, furniture, and connectivity that the firm once supplied as constant capital are now supplied by the worker, and the value of labour power does not automatically rise to cover them. The working day loses its physical boundary, and surplus labour can expand through the message answered at night and the task finished after the nominal close. Against those gains to the firm, the worker recovers the commuting time that was always an unpaid annex of the working day and may convert part of it into genuine free time. Which effect dominates for a given workforce depends on hours actually worked and costs actually borne, and must be investigated rather than assumed from the location of the desk.