WP-135 · 4 October 2026
ECONOMICS · POLICY

Who Owns the Machines

A sequel to WP-134. Part I is satire in a conjured Swift’s voice; Part II splits the question of who owns the machines and tests each part by derivation or by the passages read in Piketty, Scott and Freyre.

Satire in Part I · evidence in Part II · English and Português

Part I. A Modest Proposal for Settling Who Owns the Machines, by Entering Every Subject in the Register as Proprietor of an Equal Part

Satire. The narrator is conjured, once more, from the Deanery of St Patrick’s. His voice is not ours; he is the target. Part II is the record, with every figure traced to its source and every derivation shown.

It is a melancholy object to those who walk through this great Town, to see the engines at their work without rest, the looms that weave themselves, the cabinets that reckon without a clerk, and to observe that the labourer who once stood beside each is gone to the door with his children, to inquire whose they are. I have considered the question these many years, and am happy to report it settled. The engines belong to whoever is written down as owning them. The Register, being a book, cannot err; and a thing that is in a book is a thing that is so.

I have also made inquiry into who is written down. By the computation of an eminent Frenchman, the yield of a stock is its return multiplied by its quantity, which no honest person disputes; and the stock is held, in the best-governed Kingdoms of the present age, by one tenth of the people, in the proportion of six parts in ten or seven, while the lower half of the people hold a twentieth. I am assured this is a handsome provision for persons who have no occasion to use an engine, and that they are the happier for being spared the care of one.

I am aware that in the Plantations the same arrangement was long maintained with great regularity, and that the owner there was observed to feed his labourer well, the labourer being, in his own phrase, his engine. I think no higher regard can be shown by a proprietor to a machine, and I recommend it to the machines. I do not propose that any man be owned. I propose only that the deeds be kept in good order.

My proposal is this. Let every subject be entered in the Register as proprietor of an equal part of all the engines, so that the lower half of the people, who now hold a twentieth, shall hold a half; and let a deed be drawn for each. But as the Register must then enter some eight thousand millions of names, and the clerk is but one, I recommend that the old method be retained, whereby it enters a single name and the rest enjoy the satisfaction of being legible. I have no interest in the matter, being dead, and having never owned an engine, nor a deed.

Part II. The record

Abstract

Following WP-131, WP-132, WP-133 and WP-134, this paper asks who owns the machines and tests what three books say that bears on it: Piketty on who owns capital (1), Scott on how states make property legible (2), and Freyre on the Brazilian plantation as a system of ownership (3). We split the question into seven claims. By the accounting identity α = r × β (1, p. 52), every ten points of national income that become the yield of machines add 2.0 years of national income to their value at a 5 per cent return (our application). Piketty’s Table 7.2 puts the bottom half of the population’s share of total capital at 5 per cent in the United States in 2010 and in Europe in 1910 and 2010 (p. 248, his estimates). If machines were owned as capital is, the bottom half would receive about a tenth of its population share of machine income, and under equal shares ten times what it receives in the 2010 United States (our arithmetic; the assumption that machine ownership follows total capital is not shown). Piketty applies the same law to slave capital and finds 12 years of national income in a hypothetical society where slaves yield 60 per cent of income (pp. 162–163). Scott shows, for land and surnames, that a register creates ownership and does not only describe it (2, pp. 2–3, 35–36, 64–66); extending that to machines is ours and not shown. Freyre records the owner’s own vocabulary, the slave as capital and work machine (3, p. 107). We do not test the author’s proposal, to widen the ownership of machines. These are accounting identities, reported estimates and readings of passages, not a forecast of what automation will do.

How to read this part

Each source is numbered in order of first use and marked in the reference list with how it was read; none of the three was read cover to cover. Page numbers are printed page numbers where we could confirm them, and the limits section says where we could not. Where a figure or a derivation is ours, we say so and show it. The English renderings of Freyre are ours, and the Portuguese text needs native review.

1. The claim, split into seven

“Machines will do the work, so who owns them decides who is paid” contains seven claims. O1: income from machines accrues to whoever owns them, and the value of the stock follows from the yield by an accounting identity. O2: the ownership of capital is highly concentrated. O3: so, if machines are owned as capital is, the bottom half of the population receives about a tenth of its population share of machine income. O4: the same accounting has been applied to persons held as property, in Piketty’s treatment of slavery. O5: ownership is made by a record and not only described by it (Scott). O6: the plantation owner reckoned people as capital and as a work machine, and after abolition the mill appears where the household had been (Freyre). O7 is the author’s proposal, to widen the ownership of machines, which is a proposal and is not tested here. A companion claim, that automation will raise capital’s share of income, is not one of the seven and is not tested.

2. The identity (O1)

Piketty defines the capital/income ratio β, the share of capital income in national income α and the average return r, and states α = r × β (1, p. 52). He calls it a pure accounting identity, true by definition, and useful because it links the three quantities (p. 52). Solved for the stock, β = α / r. If an asset class yields a share α of national income, its market value is α / r years of national income; and if automation moved Δα points of national income into the yield of machines, with r unchanged, the value of the machines would rise by Δβ = Δα / r. At r = 5 per cent every ten points add 2.0 years. The identity is Piketty’s; the application to machines and the table are ours.

Return r+10 points+20 points+30 points
3%3.336.6710.00
5%2.004.006.00
7%1.432.864.29

For comparison, Piketty applies the same law to a hypothetical society in which 40 per cent of national income is the yield of land and other capital and 60 per cent is the yield of slaves to their owners, at a 5 per cent return: the capital is worth 8 years of national income, the slaves 12, the two together 20 (pp. 162–163; Figure B marks the first two). Whether automation will raise capital’s share is a different question, which the identity does not answer. Piketty treats an entirely robotised economy, in which output rises by adding capital alone, as an extreme case and says that neither extreme is really relevant (p. 217); he finds many uses for ever more sophisticated robots without needing that extreme (p. 221). We therefore do not claim that α will rise.

r = 3%r = 5%r = 7%
051015200%10%20%30%40%50%60%r = 3%20.0r = 5%12.0r = 7%8.6Piketty: yield 40% at 5% is worth 8 yearsPiketty: yield 60% at 5% is worth 12 years
Figure B. Value of an asset class, in years of national income, against the share of national income it yields. Lines for returns of 3, 5 and 7 per cent. The two rings are Piketty’s examples at 5 per cent: capital yielding 40 per cent is worth 8 years, slaves yielding 60 per cent are worth 12 (reference 1, pp. 162–163). Our arithmetic: β = α / r. Horizontal axis: share of national income yielded; vertical axis: years of national income.

3. Who owns capital (O2, O3)

Piketty’s Table 7.2 gives the shares of total capital owned by groups of the population in stylised societies (1, p. 248). In the United States in 2010 the top 10 per cent own 70 per cent (35 for the top 1 per cent), the middle 40 per cent own 25 and the bottom 50 per cent own 5; in Europe in 2010 the figures are 60 (25), 35 and 5; in Europe in 1910, 90 (50), 5 and 5 (Figure A). In the text he says that the poorer half of the population has barely 5 per cent of total wealth in 2010, as in 1910 (p. 261), and that to his knowledge no society has existed in which the poorest half owned a significant share, say a fifth to a quarter (p. 258); the low-inequality column of the table is a virtual example, never observed.

bottom 50%middle 40%next 9%top 1%
0%25%50%75%100%Europe 1910Europe 1910, bottom 50%: 5%Europe 1910, middle 40%: 5%Europe 1910, next 9%: 40%40%Europe 1910, top 1%: 50%50%US 2010US 2010, bottom 50%: 5%US 2010, middle 40%: 25%25%US 2010, next 9%: 35%35%US 2010, top 1%: 35%35%Europe 2010Europe 2010, bottom 50%: 5%Europe 2010, middle 40%: 35%35%Europe 2010, next 9%: 35%35%Europe 2010, top 1%: 25%25%Scandinavia 1970s–80sScandinavia 1970s–80s, bottom 50%: 10%10%Scandinavia 1970s–80s, middle 40%: 40%40%Scandinavia 1970s–80s, next 9%: 30%30%Scandinavia 1970s–80s, top 1%: 20%20%Low inequality (never observed)Low inequality (never observed), bottom 50%: 25%25%Low inequality (never observed), middle 40%: 45%45%Low inequality (never observed), next 9%: 20%20%Low inequality (never observed), top 1%: 10%10%Equal shares (ours)Equal shares (ours), bottom 50%: 50%50%Equal shares (ours), middle 40%: 40%40%Equal shares (ours), next 9%: 9%9%Equal shares (ours), top 1%: 1%
Figure A. Share of total capital owned by each part of the population, in five stylised societies and in an equal-shares case. Columns 1 to 5 are the shares in Piketty’s Table 7.2 (reference 1, p. 248); the equal-shares row is our own, with each group owning its share of the population. Total capital, not machines.

Dividing each group’s share by its share of the population gives machine income per person relative to the average, if machines were owned as capital is (O3, our arithmetic).

GroupEurope 1910US 2010Equal shares
top 1%50.0035.001.00
next 9%4.443.891.00
middle 40%0.130.631.00
bottom 50%0.100.101.00

In the 2010 United States column the bottom half receives 0.10 of the average per person, a tenth, and the top 1 per cent receives 35 times the average. Equal shares raise the bottom half’s share from 5 to 50 per cent, a factor of ten, with total machine income unchanged.

Two cautions. The table is for total capital, which in Piketty’s account includes housing, financial assets and business capital, and the composition changes with wealth: nearly everyone in the top decile owns a home, but real estate matters less higher up (p. 260). Machines could be held more or less concentratedly than capital as a whole; we have no data on that, so O3 is conditional on the assumption. The shares are also his stylised orders of magnitude (p. 261), not exact measurements.

4. The precedent of slave capital (O4)

Piketty counts slavery in the history of capital: a slaveholder can fully own another person’s labour capacity and offspring, and slaves were traded, inherited and counted in a slaveholder’s wealth (1, p. 46); he excludes human capital from his own definition of capital in societies without slavery because it cannot be exchanged on a market (p. 49). Applying the first law to slave capital, as in section 2, gives 12 years of national income when slaves yield 60 per cent of income at a 5 per cent return. The actual figure for the United States in 1770–1810 was about one and a half years, because slaves were 20 per cent of the population, their productivity was slightly below that of free labour and the return was nearer 7 or 8 per cent (p. 163). He also reports a typical price of ten to twelve years of an equivalent free worker’s wages, and in 1860 about $2,000 for a prime-age male slave against about $200 for a free farm labourer’s wage (p. 163). We use this for one point only: the first law is indifferent to what the owned thing is, so who owns it is decided by the record, section 5. We do not equate machines with persons, and nothing here says what a machine is owed.

5. Ownership made by record (O5)

Scott’s argument, in the passages read, is that a state makes a society legible so that it can tax, conscript and prevent rebellion, and that the records it makes are not only descriptions: a cadastral map made to designate taxable property-holders does not merely describe a system of land tenure, it creates one through its power to give its categories the force of law (2, pp. 2–3). Modern freehold is a simplification: land owned by a legal individual with wide powers of use, inheritance or sale, represented by a uniform deed of title enforced by the state’s courts and police (p. 36), and readily decipherable only to those who can read the statutes (p. 35). Permanent surnames were a project of the same kind: in China, and in the Florentine catasto of 1427, state initiative created surnames rather than recording them (pp. 64–66). On the plantation, Scott reports that colonial states preferred it to smallholdings not for efficiency but as a unit of taxation, labour discipline and political control (pp. 189–190). We searched the text for machines and automation and found no passage on who owns machines, only Le Corbusier’s machine age and craft knowledge. So O5 is read for land, names and plantations; its application to machines is our extension, and it implies one thing only: a claim to machines is as clear as the register that holds it.

6. The plantation and the mill (O6)

Freyre, in the passage read, says that the engenho owners learned that a slave driven beyond his strength yielded less than one kept in good condition, and that it was in the owner’s interest to conserve in the slave “seu capital, sua máquina de trabalho” (his capital, his work machine); he calls this a Brazilian imitation of Taylorism (3, p. 107). That is Freyre’s report of the owner’s reckoning and not a claim that the enslaved were well treated. The preface to the edition itself warns that his descriptions are at times romanced and even distorted (Cardoso, “Um livro perene”; page not confirmed), so we use him only for the owner’s vocabulary. In his own preface he also notes that, with the development of the sugar mills (usinas), the latifundium kept advancing, and that after the end of the patriarchal order in 1888 the slave was replaced by the “pária de usina” (the mill pariah) and the senzala by the mucambo (p. 39). That is one case of an ownership form changing around a machine while the concentration of land continued. Whether it is a precedent for the machines of today is not something the passages show.

7. The proposal (O7)

The author’s proposal is to widen the ownership of machines. We do not test it. The arithmetic of section 3 shows only what equal shares would change in the division of machine income, from 5 to 50 per cent for the bottom half; it says nothing about how shares would be created, held, traded or kept equal, nor about what happens to the machines’ output. Piketty, against the divergence he describes, names a progressive global tax on capital and adds that it would need a considerable degree of international coordination and that actual responses will probably be more modest (1, p. 27); we did not evaluate that either. Scott gives one practical warning: whichever proposal is chosen, someone must keep the register, and the register decides.

8. Status of each claim

ClaimStatus
O1. Machine income goes to owners; stock value = yield / returnHolds as an accounting identity (Piketty, p. 52); the application Δβ = Δα / r is ours; it says nothing about whether α rises
O2. Ownership of capital is highly concentratedReported by Piketty (Table 7.2, p. 248; text pp. 258–261); his estimates, not ours; total capital, not machines
O3. Bottom half receives about a tenth of its population share of machine incomeOur arithmetic from his shares (0.10 in the 2010 United States); conditional on machine ownership following total capital, which is not shown
O4. The same accounting was applied to persons held as propertyRead in Piketty (pp. 46, 49, 162–163); formal point only; machines are not equated with persons
O5. Ownership is made by a recordRead in Scott for land, surnames and plantations (pp. 2–3, 35–36, 64–66, 189–190); for machines it is our extension and is not shown
O6. The owner reckoned people as capital and a work machine; the mill followed the householdRead in Freyre (pp. 39, 107; pages from the edition’s index); owner’s vocabulary only; his welfare claims are not used
O7. Widen the ownership of machinesThe author’s proposal; no mechanism given; not tested
Automation will raise capital’s share of incomeNot shown; Piketty treats full robotisation as a limiting case (pp. 217, 221)
The three books were read in fullNo: passages were found by text search and read with the pages around them

9. Limits and open items

Piketty’s page numbers are the printed numbers, found by subtracting 5 from the PDF page and checked against the book’s own list of tables, which puts Table 7.2 on page 248. Scott’s are found by subtracting 16 from the PDF page and checked on running heads. The Freyre text we read has no running page numbers, so pages 107 and 39 are taken from the edition’s index and table of contents and should be confirmed against the printed book. The Piketty table is total capital in stylised societies, not machines, and the 5 per cent for the bottom half is his order of magnitude. The equal-shares row is ours and is not an observed society. Nothing here estimates automation’s effect on capital’s share; WP-134 noted the US labour share of 53.8 per cent for this question, and we did not re-read it here. The relation to WP-131 to WP-134 is one of form and theme, not proof.

References

Style: ABNT NBR 6023. Bracketed notes give how each source was read. The books are PDF copies supplied by the author and have no web address.

  1. PIKETTY, Thomas. Capital in the twenty-first century. Translated by Arthur Goldhammer. Cambridge, MA; London: The Belknap Press of Harvard University Press, 2014. PDF copy. Accessed: 4 Oct. 2026. [Read by text search and the pages around each passage cited (pp. 27, 46, 49, 52, 162–163, 217, 221, 248, 258–261); not read cover to cover; printed page = PDF page minus 5.]
  2. SCOTT, James C. Seeing like a state: how certain schemes to improve the human condition have failed. New Haven; London: Yale University Press, 1998. PDF copy. Accessed: 4 Oct. 2026. [Read by text search and the pages around each passage cited (pp. 2–3, 35–36, 64–66, 189–190); not read cover to cover; printed page = PDF page minus 16.]
  3. FREYRE, Gilberto. Casa-grande & senzala: formação da família brasileira sob o regime da economia patriarcal. 48. ed. [S. l.]: Global Editora, 2003. PDF copy. Accessed: 4 Oct. 2026. [Read by text search and the pages around each passage cited (pp. 39, 107; Cardoso’s preface); not read cover to cover; page numbers from the edition’s index and table of contents; the text is Portuguese and the English renderings are ours.]

Satirical frame: SWIFT, Jonathan. A modest proposal for preventing the children of poor people in Ireland from being a burden to their parents or country, and for making them beneficial to the publick. Dublin, 1729 (public domain). Part I echoes its title, cadence and closing disclaimer of interest.

4 October 2026.