Productive by Definition
Economics now counts as productive whatever earns an income. It was not always so. How the line between making and taking moved, and what it does to banks, the State and medicines.
In November 2009, a year after the American government had rescued the financial system, the head of Goldman Sachs, Lloyd Blankfein, defended his staff’s pay to a conference of banking analysts. “What people fail to mention,” he said, “is that net income generated per head is a multiple of our peer average. The people of Goldman Sachs are among the most productive in the world.”[1] The economist Mariana Mazzucato opens The Value of Everything with that sentence and asks a simple question about it: how would anyone show that he was wrong?[2]
Her answer is that, in modern economics, nobody could. If the value of an activity is whatever someone pays for it, then an activity that earns a great deal is, by definition, producing a great deal. “You earn income because you are productive and you are productive because you earn income. So with a wave of a wand, the concept of unearned income vanishes.”[2] Blankfein measured his staff’s productivity by their income, and then used it to justify their income. The argument goes round in a circle, and economics has no way of stepping out of it.
The Tollkeepers asked of each large income whether it was a reward for making something or a toll on something that would have existed anyway.[a] This essay asks the question behind that one: who decides what counts as making anything at all? The answer turns out to be concrete. The line is drawn in a rulebook with numbered paragraphs, the international System of National Accounts, and it has been moved several times, each time after an argument (Figure 1). Mazzucato’s case, which this essay makes with the evidence, is that every move went the same way. Finance has been brought inside the line; the State has been kept at the door, counted only at what it costs; and the language followed the accounts.
The sterile class
The first economists were sure that some work produced wealth and some did not, and they spent a great deal of effort deciding which. In 1766 François Quesnay, physician to Louis XV and leader of the group later called the Physiocrats, divided the nation into three classes. The productive class was the one that farmed: only the land gave back more than was put into it. The landowners, the king and the church, who collected the surplus, were a second class. The third he called the sterile class: “all the citizens occupied in services and work other than agriculture”, whose expenses are paid by the other two.[3] Artisans, manufacturers and merchants were all sterile.
Adam Smith thought this absurd and said so. The Physiocrats, he wrote, tried “to degrade” the artificers, manufacturers and merchants “by the humiliating appellation of the barren or unproductive class”.[4] But Smith drew his own line, only further out. Labour was productive if it fixed itself in a “vendible commodity” that lasted and could be sold to pay for more labour: the worker in a manufactory was productive, the menial servant was not. And then, in the most quoted passage of the debate: “The sovereign, for example, with all the officers both of justice and war who serve under him, the whole army and navy, are unproductive labourers.” So were “churchmen, lawyers, physicians, men of letters of all kinds; players, buffoons, musicians, opera-singers, opera-dancers, etc.”[4]
Smith is often read as calling these people parasites. He did not. The servant’s labour, he wrote in the same paragraph, “has its value, and deserves its reward”, and the sovereign’s service may be “honourable”, “useful” or “necessary”.[4] His concern was accumulation. A country grows richer when more of its income goes to work that leaves something behind to work with next year, and poorer when more goes to work that “perishes in the very instant of its production”. The distinction was a theory of growth, not a moral ranking.
Karl Marx moved the line again, and made it depend not on what the work was but on whom it was for. “That labourer alone is productive, who produces surplus-value for the capitalist.” A schoolmaster is productive when, “in addition to belabouring the heads of his scholars, he works like a horse to enrich the school proprietor”; that the owner has put his capital “in a teaching factory, instead of in a sausage factory, does not alter the relation”.[5] The same singer was unproductive when she sang on her own account and productive when an entrepreneur hired her to make money. And some activities stayed outside whatever their employer. Buying and selling, Marx wrote, “creates no value, any more than the work performed in a judicial proceeding increases the value of the subject matter of the suit”; commercial and money-dealing capital “creates neither value nor surplus-value” but takes a share of what is produced elsewhere.[5] Transport, on the other hand, he counted as production: moving a thing to where it is needed changes it.
Three theories, three lines, and they disagree about almost everything. What they share is the conviction that the line matters, because it decides who is living off whom. If the farmers produce the surplus, the rest of the nation lives on it; if manufacture does, the court and the army are a charge on it; if surplus comes from labour employed by capital, the merchant and the banker are paid out of it. Each line was also, openly, an argument about policy: Quesnay’s for a single tax on land, Smith’s for frugal government, Marx’s for revolution.
Productive because it is valued
At the end of the nineteenth century the question was abandoned. William Stanley Jevons opened his Theory of Political Economy in 1871 with what he called “the somewhat novel opinion, that value depends entirely upon utility”. Labour did not create value; at most it was a cost. “Labour once spent has no influence on the future value of any article: it is gone and lost for ever. In commerce, by-gones are for ever by-gones.”[6] Alfred Marshall, whose textbook taught economics to the English-speaking world for half a century, dismissed the old distinction directly: “It is sometimes said that traders do not produce: that while the cabinet-maker produces furniture, the furniture-dealer merely sells what is already produced. But there is no scientific foundation for this distinction. They both produce utilities, and neither of them can do more.”[7]
Lionel Robbins drew the conclusion in 1932, quoting Smith’s list of opera singers and dancers. Modern economists had rejected Smith’s idea of productivity. “So long as it is the object of demand, whether privately or collectively formulated, the labour of the opera singers and dancers must be regarded as ‘productive’. … It is productive because it is valued.”[8] In a footnote he added, with some honesty, that “it is even arguable that the reaction has gone too far”.
This is the hinge of Mazzucato’s book. Once productive means valued, and valued means paid for, the line no longer runs between activities. Everything sold in a market is inside, and only what is not sold remains outside. “What replaced it was the notion that it is only whatever fetches a price in the market (legally) that can be termed productive activity.”[2] Two things disappeared with the line. The first was rent as a category: to the classical economists it was unearned income, the landlord’s share of a surplus he did nothing to create; to the new economics it became “an imperfection that can be competed away”.[2] The second was the means to say of any activity that it earns more than it adds. If a price measures value, a high price measures high value, and the banker and the opera singer are productive for the same reason: someone pays them.
Mazzucato also suggests a motive. The shift from “value determining price to price determining value”, she writes, coincided with the rise of socialist movements that had used the labour theory of value as a weapon, and owners had reason to welcome a theory that did without it.[2] That is her reading of the history, not an established fact, and nothing in the argument depends on it. What matters is what the new theory did, not why it was adopted.
How finance became productive
Theories do not pay anyone. Statistics do, because governments steer by them. When the first national accounts were built in the 1930s and 1940s, the people who built them had to decide, in practice, where the line ran, and they knew it was a decision.[b] Simon Kuznets, who produced the first official estimates of American national income, wrote that if needs were judged by science and ethics, “we would exclude from national income many commodities and services now included. Many foods and drugs are worthless by scientific standards of nutrition and medication.” He noted that the accounts did not subtract the costs of “blatant advertising and the ruthless despoiling and defacing of the countryside”, and that unpaid work in the home, worth by his estimate more than a quarter of national income in 1929, was left out because it was not sold.[9][c] He made those choices, and he said they were choices.
Banks were a problem from the start. A bank’s main income is not a fee but the difference between the interest it receives on loans and the interest it pays on deposits. If banks were treated like any other firm, with interest counted as a transfer of income rather than a payment for a service, they would show, in the words of the first United Nations rulebook in 1953, “a negligible or even negative contribution to domestic product”.[10] Statisticians found this unacceptable, and invented a service that nobody had bought. The 1953 rules imputed a service charge to depositors. The 1968 rules said plainly why: “If the transactions of banks and similar financial institutions were treated like the transactions of other industries, their operating surplus, and perhaps their value added, would therefore be negative.” To avoid this “anomaly”, the interest margin was to be counted as a service charge. But because nobody could say who consumed it, it was charged to a “nominal industry” with an equal negative surplus, so that the banks showed a positive value added and the economy as a whole was unchanged.[10] For decades most European accounts kept the banks’ implicit services out of GDP that way.
The 1993 rules gave the service a name, financial intermediation services indirectly measured, FISIM, and asked countries to allocate it to the households and firms that use it, which adds it to GDP. The European Union made the allocation compulsory from 1 January 2005, and the 2008 rules removed the option of not doing it.[11] The interest margin had become production. The geographer Brett Christophers, who traced this history, called it “making finance productive”: by placing different activities on different sides of the production boundary, “national income statisticians effectively dictate what counts as productive”.[12]
The method has a curious consequence, which Andrew Haldane of the Bank of England pointed out in 2010 and which this blog has described before.[d] The service is measured against a reference rate meant to be free of risk. Whatever a bank earns above it on its loans counts as output, including the extra it charges for lending to riskier borrowers. So when risk rises, measured output rises: British finance reached 9 per cent of the economy’s output at the end of 2008, as the crisis broke. And the same risk counts differently depending on who bears it. “Measured financial intermediation output increases if a bank bears the risk of lending to a company. But gross output is unchanged if a household holds a bond issued by the same company.”[13] A study for the euro area estimated that removing the reward for risk would cut FISIM by up to 40 per cent.[13] The 2008 rules conceded that the assumption behind the method “has been queried”. The 2025 rules, adopted by the United Nations in March of that year, renamed FISIM, counted the bank’s work of turning short deposits into long loans, with its risks, explicitly as a service, and on the risk of default said only that there is “conceptual merit” in excluding it, without requiring anyone to do so.[14]
None of this is a conspiracy, and the accounts are not wrong by their own rules. But the rules were written, rewritten and argued over, and the effect of the rewriting is visible (Figure 2). In the United States, finance and insurance were 2.4 per cent of GDP in 1947 and 8.7 per cent in 2025.[15] Part of that rise is real: more people have bank accounts, pensions and insurance. Part of it is that the rich world decided to count the interest margin as a product, and to count the reward for risk as part of the product.
A year of record productivity
Portugal provided a clean experiment in 2023. In December 2022, 90 per cent of the stock of Portuguese mortgages carried a variable rate tied to Euribor, according to the Banco de Portugal, so when the European Central Bank raised its rates, households’ payments rose within months.[16] What banks paid on deposits barely moved. In 2023 Portuguese banks paid households 0.01 per cent a year on overnight deposits, against 0.24 per cent in the euro area; on new term deposits they paid 1.8 per cent, against 2.7 per cent.[17] The five largest banks reported profits of €4.4 billion.[17][e]
The national accounts recorded what happened as production. The value added of financial and insurance activities rose from €11.4 billion in 2022 to €16.6 billion in 2023, an increase of 45 per cent in one year, and finance’s share of the economy’s value added jumped from 5.4 to 7.0 per cent (Figure 3).[15] FISIM on loans, the implicit service charged to borrowers, nearly doubled, from €4.9 billion to €9.1 billion.[18] The sector employed 85,000 people, a thousand more than the year before. By Lloyd Blankfein’s yardstick, value added per head, Portuguese finance workers became 43 per cent more productive in a single year, rising from €136,000 to €194,000 each, while doing what they had done the year before.
Statisticians deserve credit here, and the figure shows why. In volume, the sector’s output grew by 1 per cent in 2023; the rest of the rise was treated as an increase in the price of financial services, of about 43 per cent. That is the honest answer: borrowers paid much more for much the same service. But it is an answer that shows up only in a column of price indices. In the accounts of national income, the €5 billion is the banks’ value added, the income they earned by producing something, and not a transfer from borrowers to lenders. A household whose mortgage payment went up is recorded as buying a more expensive service.
Where the money went is plain in the accounts of the sector itself (Figure 4, below). In 2019 Portuguese financial corporations paid 47 per cent of their net value added to their employees and kept 42 per cent as net operating surplus. In 2024 they paid 30 per cent to employees and kept 62 per cent.[18] The rise in measured output was the rise in profit. This is the circle Mazzucato describes, closed in a single year: the income is recorded as output because it is income, and the output then makes the sector look productive.
Counted at cost
The other half of the argument is about the State, and here the rule is written down in so many words. Government services are not sold, so they have no price, and the accounts value them at what they cost to produce: the wages paid, the materials bought, the wear on buildings and equipment. The 1993 rules then added: “The net operating surplus on the production of non-market goods or services produced by government units and NPISHs is assumed always to be zero.” The 2008 rules repeated it: “By convention, no net return to capital is included for non-market production.”[11] The European system, which is law in Portugal, prescribes the same sum of costs.[18]
The result is visible in any year of the Portuguese accounts. In 2024 the general government’s value added was €37.1 billion: €30.3 billion of pay, €7.0 billion of depreciation, and a net operating surplus of €0.15 billion, from the few things public bodies sell (Figure 4).[18] A teacher’s contribution to the national product is her salary. A hospital that cures more patients with the same staff produces, in the accounts, exactly as much as before. For years British statisticians measured whole public services that way, “output equals cost of inputs”, so that productivity could not rise by definition; in 2023 defence, policing and the administration of social security were still measured so.[19] A review of the British method in 2025 put it plainly: “Using cost to value services, as Atkinson (2005) noted, leads to a failure to recognise the value added by public services.”[19]
Nobody chose this to belittle the State. Without prices there is no obvious other way, and valuing output at cost is cautious. But conventions shape how people talk. A firm’s accounts show a surplus, a return, a contribution; the State’s show only spending. A bank is said to create wealth and a school to consume it. Taxes become a “burden” on the productive economy rather than the price of things the economy could not produce without, and a public investment is scored as a cost while the same asset in private hands earns a return. Mazzucato’s point is that this vocabulary is not neutral: “Measurements are not neutral: they affect behaviour and vice versa.”[2]
On this point the statisticians have partly agreed with her. The 2025 rules end the old convention: “a return to capital used in production is now also included in the value of output for non-market producers such as governments”.[14] A rough calculation for Australia suggested it could add up to 1.3 per cent to the level of GDP.[20] For Portugal, an illustrative estimate of our own, applying the 4 per cent nominal rate proposed in draft United Nations guidance to the State’s net fixed assets of €117 billion in 2023, gives about €4.7 billion, some 1.7 per cent of GDP. The European Union will apply the change only when it revises its own system, which it has not yet done. And it fixes the capital, not the people: a teacher’s output is still her salary. The accounts will now say the State earns a return on its buildings. They will still not say what it produces.
Who made the medicine
Mazzucato’s sharpest example is a cancer treatment. In August 2017 the American regulator approved Novartis’s Kymriah, the first therapy that re-engineers a patient’s own immune cells to attack leukaemia, for children and young adults whose disease had resisted everything else. Novartis set the price at $475,000 for a single infusion and agreed with the public insurer, Medicare and Medicaid, to be paid only when a patient responded within a month, an arrangement it presented as paying for value.[21] Mazzucato’s complaint is that the value had been made by many hands: the government accepted the “value-based pricing” model “even when the taxpayer contributed $200 million dollars towards its discovery”.[2]
The details need care, and are better for it. The regulator approves medicines, not prices; the price was Novartis’s. The $200 million is an estimate by a patients’ advocacy group, repeated by nine members of Congress in a letter, for two decades of public funding of the research behind the whole family of such therapies, not for Kymriah alone.[21] The first clinical report from the University of Pennsylvania team that developed it was supported by grants from the National Cancer Institute, and also by charities.[21] The outcome-based arrangement with Medicare was quietly dropped in 2018. The price was not: by 2025 the list price had risen to $593,533.[21]
The wider pattern is firmer than the single case. A team at Bentley University traced the research behind every new medicine approved in the United States from 2010 to 2016. Research funded by the National Institutes of Health contributed to every one of the 210, through more than $100 billion of public projects, mostly on the biology of the targets the drugs act on rather than on the drugs themselves. Extending the count to 2019, they found public research behind 354 of 356 approvals, $187 billion in all.[22] Those sums paid for science that served many products, so they are not the public’s bill for any one of them. But they make Mazzucato’s point: the riskiest and earliest part of the work, the part no firm would pay for because its use was not yet known, is mostly done with public money.
The best-known case of price parted from cost had little public research in it, which makes it useful in another way. In 2011 Gilead bought Pharmasset, the small company that had developed sofosbuvir, a cure for hepatitis C, for $11.2 billion. Pharmasset had projected a price of about $36,000 a course; Gilead charged $84,000, $1,000 a pill, for a medicine that Pharmasset had calculated could be made for about $1 a pill. In its first year on the market, Gilead’s hepatitis C drugs sold $12.4 billion, more than it had paid for the company. The American Senate’s investigation found that the price was set to “maximize revenue”; the only direct federal grant it found behind the drug was $244,479.[23] What set the price was not the cost of research, public or private. It was what a patient, or a health service, would pay rather than let someone die.
Portugal met that logic in its own parliament. On 4 February 2015, at a hearing of the health committee, a patient named José Carlos Saldanha interrupted the health minister, Paulo Macedo: “Não me deixe morrer”, do not let me die. Two days later the government announced an agreement with Gilead to treat every patient who needed it, at a price per patient that was never published; the press reported it as under €25,000, against €42,000 on the table.[24] By July 2026 the medicines authority had registered 39,519 treatments; of the patients whose results were reported, 22,409 were cured and 761 were not.[24] It is one of the Portuguese health service’s great achievements, and it was bought at a price that bore no relation to what the cure cost to make.
Where the money goes afterwards is documented too. William Lazonick and colleagues followed the eighteen drug companies in the S&P 500 index from 2006 to 2015. They paid out 99 per cent of their profits to shareholders, half as share buybacks and half as dividends. The buybacks alone came to $261 billion, 56 per cent of what the same companies spent on research and development.[25] The popular claim that drug companies spend more on buybacks than on research is false; buybacks and dividends together come close to the whole research budget.
In Portugal the State paid for most of the country’s research until recently (Figure 5). In 1995 government money funded 65 per cent of a small total; in 2024 it funded 33 per cent of a total ten times larger, €1.6 billion of €5.0 billion, slightly more than the European average.[26] Business now pays for most of it, which is what every government has wanted. The question Mazzucato asks is what the public gets back from its third. The answer, in the accounts, is nothing that shows: the grants are spending, the university salaries are costs, and the returns, when there are any, appear in the profits of whoever holds the patent.
The case for prices
The strongest reply to all of this is that prices are the only measure of value that is not somebody’s opinion. Jamie Whyte, reviewing the book for the free-market Institute of Economic Affairs, argued that Mazzucato misreads the economics she attacks. In marginal theory value determines price, not the reverse: a price settles where what the last buyer will pay meets what the last seller will accept, and “if you pay £6 for a Big Mac meal, that is because it is worth at least that much to you.” Her lines between makers and takers, he wrote, “are entailed by no theory of value; they are the personal judgements of Mazzucato about who is deserving and who not. … It is, of course, Mazzucato who is the beholder.”[27] Friedrich Hayek’s older argument stands behind this: prices carry information that no single mind could collect, about scarcities and wants spread across millions of people.[28]
Other critics are more sympathetic but still sceptical. The economist Diane Coyle, author of a short history of GDP, found Mazzucato’s account of the production boundary “confusing”, and pointed out that the official boundary is a statistical convention that statisticians themselves criticise, not a hidden ideology.[29] Thomas Hemphill, writing in the Cato Journal, called her view of venture capitalists “strangely naive”, since most of their investments fail, and noted that the price of hepatitis C cures fell because of competition, not because of any theory of value.[30] A study for the Tufts centre that tracks drug development, sponsored in part by Pfizer, found that among 26 of the most transformational medicines, public science led most of the basic discoveries, while private firms led most of the discovery and development of the drugs themselves.[31] And the book has errors: it describes Quesnay’s sterile class as landlords, nobles and clergy, which Quesnay did not, and its $200 million is looser than its prose.
These points are fair, and they narrow the argument without defeating it. Prices are excellent signals inside a market, and nothing here is a case for doing without them. But the argument is not about the price of a hamburger. It is about where the line is drawn in the accounts, and the accounts are full of judgements that are not prices: the imputed service of the banks, the reward for risk counted as output, the State valued at cost, the 2025 decision to add a return on public capital. Kuznets asked of the people who draw the line, “why should we accept the judgments of this particular group of people concerning productivity and economic value?”[9] The question cuts both ways. If Mazzucato’s line is a judgement, so is the one already in use. The choice is not between a judgement and none, but between judgements made in the open and judgements buried in paragraph 6.166.
Mazzucato herself does not want a new wall between virtuous and parasitic trades. “The point is not to create a stark divide, labelling some activities as productive and categorizing others as unproductive rent-seeking.”[2] The case needs no labour theory of value. It needs only the question that Smith and Marx asked and modern economics stopped asking, applied case by case: is this income a payment for something made, or a share of something made elsewhere? The Portuguese banks in 2023 show that the question can be answered with the statistics that already exist.
Redrawing the line
If the line is a decision, it can be decided better. None of what follows needs a planned economy or a new theory of value. It needs the decisions to be made in the open.
Publish the toll next to the service. The 2025 rules already ask any country that removes the reward for default risk from banks’ output to publish the other version too. Portugal’s statistics institute could publish both, every year, and show finance’s value added in volume beside its value in euros. A year like 2023 would then read as what it was: a jump in the price of the same service.
Count the State’s output, not only its costs. Apply the 2025 return on public capital as soon as the European system allows, and go further where it can be done: measure what schools, hospitals and courts achieve, as British statisticians have done for health and education since the Atkinson review. The warning in When the Score Becomes the Goal applies: a measure that becomes a target gets gamed. That is a reason to measure several things, not to keep measuring only the wage bill.
Attach conditions to public research money. When the State pays for the riskiest stage of an invention, it can ask for something in return: a price commitment, access for the health service, a share of the royalties, or a stake. Mazzucato’s institute at University College London has set out such conditions for medicines.[32] The American law that lets the government license a publicly funded invention to others when it is not made available on reasonable terms has never once been used in more than four decades, and the 2023 attempt to say that price could be a reason is still a draft.[32] A condition that is never enforced is not a condition. Portugal, which spends €1.6 billion of public money a year on research and plans €688 million through its science foundation in 2026, could write conditions into its grants without waiting for anyone.[26]
Change the words. Stop describing public services as spending and private ones as wealth creation. Smith, who classed the sovereign and the army as unproductive, still wrote that their labour “has its value, and deserves its reward”.[4] That was more generous than much of what is said today.
The objections deserve their due. Conditions on research grants can drive private partners away, and a firm that takes most of the development risk is entitled to most of the return. Measures of public output are hard to build and easy to game. A line drawn by officials can be captured, like any other. These are reasons to draw the line carefully, in public, and to revise it when it fails. They are not reasons to leave it where it is, drawn by default, in favour of whoever is paid the most.
A decision, not a fact
Lloyd Blankfein was right, by the definition then in force. His staff earned a great deal per head, the accounts counted what they earned as what they produced, and so they were among the most productive people in the world. The trouble was the definition. Quesnay drew the line around the farm, Smith around the workshop, Marx around the capitalist’s employ, Robbins around anything with a price; the statisticians moved it in 1953, 1968, 1993, 2008 and 2025. Each move was argued, written down and adopted by people who could have chosen otherwise. It will move again. The only question is whether the people whose incomes depend on it will be the ones who decide where.
This piece was written collaboratively with Claude Opus 5.5 (Anthropic): human specification, editorial direction and critical review; machine research, analysis and drafting. It starts from M. Mazzucato, “Takers and Makers”, as republished by Evonomics in June 2019 from The Value of Everything (2018), and argues her case; counter-evidence is stated where it exists. Classical texts are quoted from the original editions; the translation of Quesnay is our own. The rules for national accounts are quoted from the United Nations manuals (the 2025 text from its pre-edit version, whose paragraph numbers may change). Where Mazzucato’s examples were loose (the Kymriah funding, Quesnay’s classes) the text gives the documented version. Portuguese hepatitis C prices come from press reports, since the agreed price was never published. The estimate of a return on Portuguese public capital is the author’s own and is illustrative. The figures are drawn by scripts/productive.py from published tables (numbers in docs/productive-results.json, figures in docs/productive-figures.html). The cover is “Arcadas na Praça do Comércio” by Bert K., Wikimedia Commons, CC BY 2.0, cropped.
Authored by: Luis Matos Ferreira — Physicist, Developer, Writer
Ideas this essay uses without stopping to explain them, and where this blog does. The letters in the text point here; the links go straight to the section.
- The Tollkeepers — rent and reward, after Stiglitz, and Portugal’s rendas excessivas.
- Earned and Unearned — how work and ownership share income in Portugal.
- Growth of What and Worth Producing — what GDP measures, and what is worth producing.
- The State’s Ledger — where the Portuguese State’s money comes from and where it goes.
- Bullshit jobs — on work that even those who do it think pointless.
- G. Farrell, “Blankfein defends pay levels for ‘more productive’ Goldman staff”, Financial Times, 11 November 2009, quoted in full by H. Blodget, Business Insider, 11 November 2009; remarks at the Bank of America Merrill Lynch Banking and Financial Services Conference, New York, 10 November 2009.
- M. Mazzucato, “Takers and Makers: Who are the Real Value Creators?”, Evonomics, 30 June 2019, adapted from The Value of Everything: Making and Taking in the Global Economy, Allen Lane, 2018; the book, Introduction, chapter 1 and chapter 2 (“The Production Boundary Becomes Malleable”, “The Disappearance of Rent and Why It Matters”).
- F. Quesnay, “Analyse de la formule arithmétique du Tableau économique”, Journal de l’agriculture, du commerce et des finances, June 1766; Wikisource; translation by the author.
- A. Smith, An Inquiry into the Nature and Causes of the Wealth of Nations (1776), Book II, chapter III, paragraphs 1–2, and Book IV, chapter IX; Project Gutenberg.
- K. Marx, Capital, vol. I (1867), chapter 16; vol. II (1885), chapter 6; vol. III (1894), chapter 17; Theories of Surplus-Value, chapter IV and addenda.
- W. S. Jevons, The Theory of Political Economy, Macmillan, 1871, pp. 2 and 159; Internet Archive.
- A. Marshall, Principles of Economics, 5th ed., 1907, Book II, chapter III, section 1; Internet Archive.
- L. Robbins, An Essay on the Nature and Significance of Economic Science, Macmillan, 1932, chapter I, section 2, p. 8.
- S. Kuznets, National Income: A Summary of Findings, NBER, 1946, Part IV, pp. 122–127; National Income and Its Composition, 1919–1938, NBER, 1941, vol. I, p. 51.
- United Nations, A System of National Accounts and Supporting Tables, 1953, chapter V, p. 32; A System of National Accounts, 1968, paragraphs 6.32–6.35.
- United Nations et al., System of National Accounts 1993, paragraphs 6.91 and 6.124–6.131; System of National Accounts 2008, paragraphs 6.94, 6.130, 6.163–6.166, 20.71, A3.27 and A4.33; Commission Regulation (EC) No 1889/2002, article 2.
- B. Christophers, “Making finance productive”, Economy and Society 40(1), 2011, pp. 112–140 (abstract).
- A. Haldane, S. Brennan and V. Madouros, “What is the contribution of the financial sector: miracle or mirage?”, in The Future of Finance: The LSE Report, 2010; J. Burgess, “Measuring financial sector output and its contribution to UK GDP”, Bank of England Quarterly Bulletin, 2011 Q3, citing A. Colangelo and R. Inklaar (2010).
- United Nations et al., System of National Accounts 2025, pre-edit version, overview and paragraphs 7.103, 7.183–7.185, A4.32–A4.34; UN Statistical Commission, decision 56/101, March 2025.
- US Bureau of Economic Analysis, GDP by industry, historical data 1947–1997 and table “Value Added by Industry as a Percentage of Gross Domestic Product”, September 2026; Eurostat, nama_10_a10 and nama_10_a10_e, retrieved 16 October 2026.
- Banco de Portugal, Relatório de Estabilidade Financeira, May 2023, Caixa 1.
- European Central Bank, MFI interest rate statistics, households, Portugal and euro area, annual averages for 2023; Lusa compilation of the five largest banks’ results, “Ano excecional dos lucros da banca em 2023”.
- Eurostat, non-financial transactions by sector (nasa_10_nf_tr; FISIM computed as interest before allocation less interest after it) and government finance statistics (gov_10a_main), retrieved 16 October 2026; European System of Accounts 2010 (Regulation (EU) No 549/2013), paragraph 3.49.
- J. Heys, “Measuring public service productivity”, ESCoE blog, 29 July 2019; UK Statistics Authority, National Statistician’s Independent Review of the Measurement of Public Services Productivity, March 2025, Table 1.
- UN Advisory Expert Group on National Accounts, “Sum of costs”, 2023; ISWGNA, “Guidelines for estimating the net return to capital for non-market production”, draft for global consultation; Eurostat, nama_10_nfa_bs.
- Novartis, “Novartis receives first ever FDA approval for a CAR-T cell therapy, Kymriah”, 30 August 2017; STAT, “A $475,000 price tag for a new cancer drug”, 31 August 2017; letter of Representatives L. Doggett and others to the CMS Administrator, 13 September 2017; D. L. Porter et al., New England Journal of Medicine 365, 2011, pp. 725–733 (grant list in PubMed 21830940); Petrie-Flom Center, 12 July 2018; Pulice and Schulman, “CAR-T therapy: escalating costs in an expanding market”, HMPI, February 2026, Table 1.
- E. Galkina Cleary et al., “Contribution of NIH funding to new drug approvals 2010–2016”, PNAS 115(10), 2018, pp. 2329–2334; E. Galkina Cleary et al., “Comparison of Research Spending on New Drug Approvals by the National Institutes of Health vs the Pharmaceutical Industry, 2010–2019”, JAMA Health Forum 4(4), 2023, e230511.
- US Senate Committee on Finance, The Price of Sovaldi and Its Impact on the U.S. Health Care System, S. Prt. 114-20, December 2015.
- Público, “Governo chega a acordo para tratar 12 mil doentes com hepatite C em três anos”, 5 February 2015, and “Hepatite C: acordo sobre medicamento prevê pagar por doente tratado”, 6 February 2015; INFARMED, Hepatite C — Monitorização dos tratamentos, 17 July 2026.
- W. Lazonick et al., “US Pharma’s Financialized Business Model”, Institute for New Economic Thinking Working Paper 60, September 2017, pp. 3–4.
- Eurostat, rd_e_gerdfund, retrieved 16 October 2026; Fundação para a Ciência e a Tecnologia, QUAR 2026, financial resources.
- J. Whyte, “‘The Value of Everything’ adds up to nothing”, Institute of Economic Affairs and CapX, July 2018.
- F. A. Hayek, “The Use of Knowledge in Society”, American Economic Review 35(4), 1945, pp. 519–530.
- D. Coyle, “Private and public value”, The Enlightened Economist, 8 May 2018.
- T. A. Hemphill, review of The Value of Everything, Cato Journal 39(3), 2019, pp. 737–743.
- R. Chakravarthy et al., “Public and Private Sector Contributions to the Research and Development of the Most Transformational Drugs in the Last 25 Years”, Tufts Center for the Study of Drug Development, January 2015, Table 2.
- UCL Institute for Innovation and Public Purpose, “The people’s prescription”, Policy Report 2018-10; US Government Accountability Office, GAO-26-107885, February 2026.
Comentários
Enviar um comentário