DEEP READING · READ 938
Le Capital au XXIe siècle · Thomas Piketty · 2013 (English translation 2014)
So long as the return on capital (r) exceeds the rate of economic growth (g), wealth accumulated in the past grows faster than income earned in the present, and a society drifts quietly back toward a world where inheriting beats working; the egalitarian decades everyone took for the normal state of affairs were in fact an exception manufactured by two world wars, a depression and confiscatory tax rates. The book's real weight is not in the inequality but behind it — three centuries of tax and estate records from twenty-odd countries, which turn inequality from a question of opinion into a question with a data series.
Piketty is an economist at the Paris School of Economics and, with Atkinson, Saez, Zucman and others, a builder of the World Inequality Database — a project that reconstructs the long-run history of distribution from tax filings and estate records rather than sample surveys. This 2013 book is the reckoning of twenty years of that data work. What made it a freak academic bestseller after the 2014 translation was not a new theory but the fact that readers could finally see a curve three centuries long: so that is what the history of inequality actually looks like.
Pin the symbols down first. r is the rate of return on capital: what the whole stock of someone's wealth — property, shares, bonds, business equity — yields in an average year, counting rent, interest, dividends and capital gains. g is the growth rate: how much bigger the economy's total output and income get each year.
Three centuries of data say r sits fairly stubbornly around 4–5%, while g was near zero for most of history, spiked to 3–4% in the postwar reconstruction and baby boom, and has since settled back to 1–1.5%. So r > g is the default state, and the mid-century period when r roughly equalled g is the aberration.
Spell out what the gap does, or it stays a slogan. Suppose average incomes rise 1% a year while an existing fortune earns 5%. Even if only part of that return is ploughed back, the fortune grows larger relative to the whole society every single year. Over three decades, anyone climbing on wages falls steadily behind — not for lack of effort, but because they are running on the slower track. Piketty's phrase for it: the past devours the future. Dead accumulation outpaces living labour.
Here is the misreading to head off: r > g is not the cause of inequality, it is an amplifier. Piketty is explicit — what the gap determines is how concentrated the steady-state distribution of wealth ends up being. A wider gap means a more extreme equilibrium; a narrower one means a milder one. Nor does it predict catastrophe, because wars, taxes, divided estates, bad investments and spendthrift heirs all push the other way. It says only this: absent any counter-force, the system's default direction is concentration, not dispersion.
The book presents two "fundamental laws of capitalism." They sound forbidding and are in fact two napkin formulas — but they are the engine of the whole argument.
β is the capital-income ratio: a country's total private wealth expressed as a number of years of national income. The second law says β tends toward s / g, where s is the savings rate. Put numbers in: a country saving 10% a year and growing at 2% converges on 10% ÷ 2% = 500%, meaning its wealth equals five years of national income. Let growth fall to 1% and the same savings rate drives β to 1000% — double.
This is the book's most counter-intuitive and most important insight: slower growth by itself mechanically increases the weight of capital. Nobody has to get greedier; no institution has to decay. The numerator (accumulated wealth) keeps compounding while the denominator (annual income) grows more slowly, so the ratio climbs. The history matches: European β stood at 600–700% before 1914, was smashed down to 200–300% after the Second World War, and is back to 500–600% today. We are living on the upswing of that curve.
α is capital's share of national income — how much of each year's output lands with the owners of capital rather than with labour. The first law, α = r × β, is a pure accounting identity rather than a theory: the rate of return times the relative size of the capital stock is the fraction capital takes. If β reaches 600% and r is 5%, then α = 30%.
Chain them together and you have the book's spine: growth slows → β rises → so long as r does not fall proportionally, α rises → and ownership of wealth is far more concentrated than wages → so overall inequality rises. That third link — "so long as r does not fall proportionally" — is the weakest, and the criticism section below pays it the attention it deserves.
| Formula | How to read it | What it says |
| β = s / g | wealth stock as years of income | slower growth makes capital heavier (10%÷2%=500%; 10%÷1%=1000%) |
| α = r × β | capital's cut of national income | an identity: bigger β, bigger cut — unless r falls in step |
| r > g | the stock outruns the flow | not a cause but an amplifier: it sets how concentrated the equilibrium is |
Three formulas carry the book: two identities and a steady state, then their consequence.
This is the one chart to remember, and the book's hardest blow to the profession.
The standard answer for decades came from Kuznets: the inverted U, in which inequality rises early in industrialisation and then falls of its own accord as labour leaves agriculture, schooling spreads and technology diffuses. It was a comforting curve, because it wrote equality into development as an automatic reward.
Piketty's rebuttal is blunt: Kuznets was looking at American data from 1913 to 1948 — the most violent thirty-five years in human history. Extend the series across three centuries and you do not get an inverted U. You get a U: very high inequality before 1914, a cliff-edge collapse between 1914 and 1945, a low plateau through the postwar decades, and a steady climb resuming around 1980.
What caused the collapse? Not market self-correction but a chain of external destruction and political decisions: physical wartime destruction, interwar hyperinflation that reduced bondholders' claims to waste paper, the bankruptcies of the Depression, the loss of colonial assets, postwar nationalisation — and above all progressive taxation. Top marginal income tax rates in the United States and Britain sat at 80–90% for decades, with estate taxes to match. That was not an accident; it was a choice a generation made under the pressure of war and crisis.
So the book's real historical claim is this: the relative equality our parents took for the normal condition of a modern society was an interlude, propped up by catastrophe and confiscatory rates; once the catastrophe receded and the rates were flattened, the old gravity resumed. That is the idea that changes how you see things: it removes the assumption that equality arrives on its own.
Schematic (shape follows the book's European series; not exact readings): the fall was produced by war, inflation and punitive tax rates, not by markets correcting themselves.
Piketty insists on separating two layers, because they are not the same order of magnitude.
Inequality of labour income is comparatively mild: in a typical rich country the top 10% of earners take roughly a quarter to a third of all wages, and the bottom 50% take roughly a quarter to three-tenths. A large gap, but the same league.
Inequality of ownership is another world entirely. In Europe today the top 10% own something like 60–70% of all wealth and the top 1% around a quarter; in the United States it is more extreme still. And in every country and every era, the bottom 50% own essentially nothing — around 5%. That is the most frequently repeated and most quietly damning number in the book. In Belle Époque Europe the top 10% held 90% of wealth and the top 1% more than 60% on their own.
Why does the distinction matter? Because it decides where policy has to aim. If inequality grows mainly out of wages, the answers are education, skills and training. If it grows mainly out of property, education can barely touch it — no amount of studying catches an asset that has been compounding for thirty years. Piketty's judgement is that the twenty-first-century battleground is shifting from the first to the second.
The book's best-known rhetorical device is reading nineteenth-century novels as data.
In Balzac's Père Goriot, the convict Vautrin advises the poor student Rastignac to give up on law: grinding his way up the profession will make him a respectable prosecutor of modest means by middle age, whereas marrying a particular heiress would hand him roughly ten times that income tomorrow. Vautrin's arithmetic was correct for his century. Jane Austen works the same way: characters are priced directly in pounds per year, and behind that annuity sits an estate or a bond yielding about 5%. Her readers needed no explanation, because r ≈ 5% was common knowledge.
That is Rastignac's dilemma: when r sits far above g, betting your life on talent and diligence is, in purely financial terms, a bad trade. The mid-twentieth century inverted the answer for the first time — war and taxation washed out old fortunes while rapid growth lifted wages — and it is that inversion which produced the life script we still hand to children: study hard, get a good job.
The data say the script is losing its footing. Piketty computes the annual inheritance flow for France — bequests and gifts each year as a share of national income. It ran at 20–25% through the nineteenth century, collapsed to around 5% in the 1950s, and is back to roughly 15% today. Inheritance is recovering something close to its old weight in the economy. Seen this way, young people's preoccupation with whether their family owns property is not a moral decline; it is the r−g gap becoming visible in private life.
Capital alone cannot explain the United States. The share of national income going to the top 10% rose from about 30–35% around 1980 to 45–50% today, and the engine of that rise was not capital gains but salaries — specifically executive pay. Piketty names this group the supermanager, to distinguish them from the rentier who lives on inherited property.
The orthodox explanation is marginal productivity: a person is paid what they add to the organisation. Piketty's objection comes in two parts. The technical part: the individual marginal product of a top executive is in practice unmeasurable — a multinational's profit is produced jointly by tens of thousands of people, by capital, brand, luck and the macroeconomic cycle, and there is no reliable way to assign a slice of it to one person. The substantive part is his real claim: what sets top pay is social norms and bargaining power, and the top marginal tax rate is the hardest anchor those norms have.
His evidence is cross-country. Line up how far each country cut its top marginal rate against how far its top income share rose, and the two track closely: the countries that cut hardest (the US, the UK) saw top shares surge; those that cut little (Germany, France, Japan) barely moved — and per-capita growth rates did not diverge to match. If high rates did not damage growth, he argues, then the pay that low rates unlocked was not newly created value so much as value transferred from somewhere else.
From this he coins a pointed phrase — meritocratic extremism: a society that no longer believes in birthright still needs a story to justify extreme rewards, so it invents the irreplaceable genius. And the specifically American risk is that both forms stack: inherited fortunes growing back, plus supermanager pay on top.
One finding often skipped over is logically crucial: r is not one number shared by everyone. The larger the fortune, the higher the realised return.
The cleanest evidence is American university endowments, whose accounts are public, consistently reported and long-running. Over 1980–2010, the very largest endowments — Harvard, Yale, Princeton — earned real returns above 10% a year, while the smallest earned around 6%. The difference comes from access: a large fund can afford elite investment staff, can hold illiquid but high-yielding assets (private equity, hedge funds, timberland, commodities), and can absorb risks that would be fatal to a small fund and are rounding errors to a big one.
The same pattern shows in the Forbes list: between 1987 and 2013 the largest fortunes grew at roughly 6–7% a year in real terms, far above the roughly 2% growth of average world wealth.
Why it matters: it means that even with no inheritance and no difference in savings behaviour, differential returns alone make wealth diverge. Compounding is already an amplifier; when the multiplier itself scales with the principal, the divergence stops being linear. This is also the direct argument for making a capital tax progressive — because the rate of return is already progressive.
Having diagnosed, he prescribes: an annual, progressive, globally coordinated tax on wealth — perhaps 0.1% on modest holdings, 1% above a million euros, 2% above several million, higher still at the very top. He concedes in the book that this is not politically achievable in the near term and calls it a "useful utopia": not a bill to be passed as written, but a yardstick against which every realistic proposal can be measured.
Crucially, the function he ranks first is not revenue but transparency. No government today genuinely knows what its wealthy own; assets sit inside cross-border structures and tax havens, and statisticians are left estimating. To levy the tax at all you would first have to build asset registries and automatic international exchange of banking information — and once wealth is visible, even a low rate changes the terms of public argument. The other half of the programme is a return to steeply progressive income tax; he estimates the revenue-maximising top rate for the United States at around 80%.
He also issues a warning from the other side: if there is neither taxation nor transparency, history's remaining methods for dealing with excess debt and concentrated wealth are inflation, protectionism and nationalist backlash — all three worse than a tax. Worth noting that in the decade since publication the world has edged in his direction: the OECD/G20 automatic exchange of financial account information has been running since 2017, and a global minimum corporate tax was agreed in 2021. Neither is his tax, but both are his kind of transparency.
Four parts, running in a straight line:
1. The return on capital has sat at 4–5% across three centuries while long-run growth manages 1–1.5%; while r exceeds g, existing wealth outruns present labour — the past devours the future.
2. State it precisely, though: r > g is not the cause of inequality but its amplifier — it sets how concentrated the equilibrium becomes, predicts no catastrophe, and can be undone at any time by taxes, divided estates and failed bets.
3. β = s / g: slower growth by itself mechanically thickens the weight of capital. Save 10% and grow at 2% and national wealth settles at five years of income; grow at 1% and it settles at ten.
4. α = r × β is an accounting identity: a thicker stock means a larger slice for capital — unless r falls in step, which is precisely the weakest link and the one critics hit hardest.
5. The long curve is not Kuznets's inverted U but a U: twentieth-century equality was manufactured by two world wars, hyperinflation and 80–90% top tax rates, not by markets healing themselves — and the curve turned back up once the shocks passed.
6. Inequality of earnings and inequality of ownership are different orders of magnitude — and in every country and every era, the bottom half of the population owns essentially nothing.
7. Vautrin told Rastignac to stop studying law and marry the heiress: when r sits far above g, betting a life on talent and diligence is financially a bad trade — and France's inheritance flow has climbed from a postwar 5% back to roughly 15% of national income.
8. The surge in the American top share was driven by executive salaries, not capital gains; Piketty attributes it to bargaining and norms rather than productivity, his evidence being that the countries that cut top tax rates hardest saw top shares rise most, with no matching gain in growth.
9. r is not the same number for everyone — bigger principal, higher return (1980–2010: over 10% real for the largest university endowments, around 6% for the smallest), so the top diverges even without inheritance, which is why a capital tax has to be progressive.
10. The line to keep: the first purpose of his global capital tax is not to collect money but to make wealth visible, since no government today truly knows what its rich own — and the book's real legacy is not the inequality r > g but the fact that it turned inequality from a clash of opinions into an argument with three centuries of data behind it.