Piketty's r > g: Why Owning Beats Working, Almost Always

The French economist Thomas Piketty did something unusual for a man who wanted to explain inequality: instead of building a model, he went digging through two centuries of tax returns. What he came back with was a book — Capital in the Twenty-First Century, published in French in 2013 and in English translation the following year — and a three-character expression that has been argued about ever since. The expression is r > g. It says that the return on capital tends to run ahead of the growth rate of the economy. And it explains more about who ends up rich than any story you have ever been told about talent, grit or getting up early.

I want to take it seriously here, because it is usually either dismissed as ideology or repeated as a slogan, and it deserves neither. It is a modest little arithmetic observation with enormous consequences, and the consequences are the point.

What r > g actually says

Two numbers. The first, r, is the average annual return on capital — what wealth earns when it is put to work. Rent from property, dividends and capital gains from shares, interest on bonds, profits from a business you own but do not personally staff. The second, g, is the growth rate of the economy — which, over time, sets the ceiling on how fast incomes from work can rise, because wages come out of national output and national output is what g measures.

Now hold them next to each other. If capital earns, say, four or five percent a year while the economy grows at one or two, then wealth compounds faster than wages. Not because the owner is cleverer or works harder. Because compounding is relentless and the exponent is bigger. A fortune left alone doubles while the salary it is compared against crawls. Over one year the gap is invisible. Over a career it is uncomfortable. Over three generations it is a different social class.

That is the whole mechanism, and its most disturbing feature is how morally empty it is. No villain is required. Nobody has to cheat, bribe, collude or exploit for the distribution to tilt. You can assume perfect markets, honest dealing and impeccable behaviour from everyone involved, and the tilt still happens, because it is a property of the arithmetic rather than of anyone's character. This is why Piketty's argument stings in a way that ordinary anti-corruption complaints do not. Corruption you can prosecute. Compound interest you cannot.

No villain is required. You can assume perfect markets and impeccable behaviour from everyone involved, and the tilt still happens — because it is a property of the arithmetic, not of anyone's character.

One caution, and Piketty is careful about it even when his popularisers are not: r > g is a tendency, not an iron law of physics. He presents the gap between the two rates as one of the important forces shaping how wealth concentrates — a force that has usually operated, not one that must operate everywhere and always. Anyone who tells you the formula guarantees inequality rises forever has flattened the claim into something its author did not make. The honest version is quieter and harder to dismiss: absent something pushing the other way, this is the direction things drift.

The method is the achievement

The reason the book landed as heavily as it did has less to do with the formula than with the evidence behind it. Economics has a long habit of reasoning about inequality from assumptions — build a model, tune the parameters, derive a result. Piketty and his collaborators did something closer to archaeology. They went into national archives and pulled out income tax records, estate and inheritance filings, property registers and wealth surveys, some running back to the eighteenth century, and assembled them into consistent long-run series for France, Britain, the United States and other countries.

This matters more than it sounds. Household surveys, the usual tool, are notoriously bad at the top — the very rich are few, hard to sample and disinclined to answer honestly. Tax and inheritance records are not perfect either, but they capture the top in a way surveys never do, and they stretch back far enough to see shapes that a decade of data would hide entirely. The formula is a summary. The archives are the contribution.

The U-shaped century

What the archives showed was a curve, and the curve is the part of the book I find hardest to argue with.

Before the First World War, wealth in Europe was staggeringly concentrated — a rentier society in the literal sense, where the top sliver of families owned most of everything and lived on the income it threw off. Then came a violent middle. Two world wars destroyed physical capital outright. The Depression wiped out fortunes. Inflation ate bondholders alive. And states, needing to pay for wars and then for reconstruction, imposed steeply progressive income and inheritance taxes of a kind that would be unthinkable now, alongside expanded public services and stronger labour bargaining.

The result was a genuine compression. For roughly the middle third of the twentieth century, the gap narrowed, a large middle class formed and owned property for the first time in history, and it became possible to believe that this was simply what a maturing economy did — that development naturally produced broad prosperity.

Then, from around 1980, the curve turned back up. Top tax rates fell, capital moved more freely, labour's bargaining position weakened, and inherited wealth began reasserting itself as a route to the top. The shape across the whole period is a U. And a U is devastating to the comforting story, because it shows that the equal middle decades were not an inevitable stage of development. They were an anomaly produced by catastrophe and by policy — a period when war, depression and deliberate taxation temporarily overwhelmed r > g. When those forces faded, the old drift resumed.

That is not a new pattern so much as an old one returning. The concentration of a surplus in the hands of whoever controls the store is about as old as settled society itself — it is there at the first granary and inequality, where the ability to hold grain across seasons created both the first accounting and the first aristocracy. What Piketty adds is the measurement: he can show you the shape of it in the tax records, year by year, rather than asking you to take it on faith.

What he proposes, and why it is hard

If the drift is arithmetic, only something deliberate can counteract it. Piketty's headline proposal is a progressive annual tax on net wealth — small rates on large fortunes, rising with size — coordinated internationally so that capital cannot simply relocate to whichever jurisdiction charges least. Alongside it he argues for far more transparency: automatic exchange of financial information between countries, public registers of who owns what, an end to the anonymity that makes hidden wealth both untaxable and unmeasurable.

He has been candid that the coordinated wealth tax is a "useful utopia" — a standard to argue toward rather than a bill anyone is about to pass. The transparency half strikes me as both more achievable and, in some ways, more radical. You cannot tax what you cannot see, but you also cannot argue about what you cannot see. A great deal of the political peace around concentrated wealth rests on nobody having the numbers.

The critics deserve a hearing

The book took serious fire, and some of it landed.

• The data handling. In May 2014 the Financial Times' economics editor, Chris Giles, went through Piketty's wealth spreadsheets and reported unexplained adjustments and transcription problems, particularly in the British series. Piketty published a detailed rebuttal within days, and most economists concluded the broad findings survived — but the specific UK wealth numbers took a real knock, and "he showed his workings and they were checkable" is a defence that cuts both ways.
• Whether r > g reliably holds. The gap is not constant. Returns on capital vary by asset, by era and by how much you own; small savers do not earn hedge fund returns. Several economists have argued the relationship does not by itself imply the wealth-to-income ratio must keep rising, since it depends on how easily capital substitutes for labour and on how much the rich consume rather than reinvest.
• How capital is measured. The sharpest technical objection came from Matthew Rognlie, who decomposed the rising capital share and found a great deal of it was housing — land prices in desirable places — rather than a broad surge across machines, factories and productive assets. If the story is substantially about property scarcity in a few cities, the remedy looks less like a global wealth tax and more like planning reform.
• Whether the tax is workable. Valuing illiquid assets annually is genuinely hard, capital is mobile, and coordination among states that compete for it has a poor historical record. Others argue institutions and politics drive inequality more than any mechanical ratio does — which is the line Acemoglu on power and progress takes, locating the cause in who holds power over technology and rules rather than in arithmetic.

I find the housing critique the most interesting and the least damaging, because it relocates the problem rather than dissolving it. Rent extracted from a scarce asset you happen to own is still income without work. That is the definition of the rentier economy, and it is the same structure wearing a different hat — which is also, incidentally, what happens when platforms and networks become the scarce asset and the toll gets collected digitally, the pattern behind how technology gets captured.

An economy where owning beats working is a distribution choice, not a law of nature. It was chosen differently once, within living memory, and the records are there to prove it.

The part that stays with me

Strip away the disputes and one thing remains standing: for most of recorded history, the surest way to be rich has been to start rich. The mid-twentieth century made us forget that, and we built an entire folklore of merit on top of the forgetting — the idea that where you end up is mostly a verdict on what you did. Piketty's archives say otherwise. They say the verdict is substantially a matter of what you inherited, and that the decades when this was less true were held open by deliberate effort and extraordinary circumstance.

Which is, in the end, a hopeful finding, though it rarely gets read that way. If concentration were a law of nature we would simply have to live with it. It is not. It is the default, and defaults can be overridden — by taxation, by transparency, by the boring institutional work of deciding that unearned income should carry more of the load than earned income does. We have overridden it before, within living memory, and we have the tax records to prove it.

An economy where owning beats working is not something that happened to us. It is something we are choosing, every year we decline to choose otherwise.