The most repeated statistic in family wealth is not a statistic. It is a proverb wearing a lab coat.
The most repeated statistic in family wealth is not a statistic. It is a proverb wearing a lab coat.
You have heard the line. Seventy percent of wealthy families lose their wealth by the second generation, ninety percent by the third. Advisers open keynotes with it. Banks build brochures around it. Family offices quote it to justify their fees. And somewhere in your own head, if you have built anything worth passing on, the line sits there like a diagnosis you never asked for: your grandchildren are statistically scheduled to blow it.
Here is the problem. The evidence for that schedule is thin to the point of vanishing, and the best long-run data we have points the other way. Advantage, it turns out, is one of the stickiest things a family can own. What evaporates in three generations is something else entirely, and once you can name it, you can build against it.
Start with the strangest fact about the three-generation curse: everyone has one.
The Americans say shirtsleeves to shirtsleeves in three generations. The Chinese say fu bu guo san dai, wealth does not pass three generations. The Japanese version is agrarian: the third generation returns to the rice paddies. The Scots put it most brutally: the father buys, the son builds, the grandchild sells, and his son begs. The Italians go from stables to stars to stables. The Brazilians and Mexicans have their own variants about rich fathers and beggar grandsons.
Advisers usually present this universality as proof. Look, every culture noticed the same pattern, so the pattern must be real. Read it again and the logic collapses. Every culture also has proverbs about lazy neighbors and dishonest merchants and mothers-in-law. Proverbs are not field data. They are compressed anxiety.
What the universality of the shirtsleeves proverb actually proves is that the fear is universal. Every wealth creator in every century has looked at a grandchild playing in the courtyard and felt the same cold thought: this child did not earn any of this, and may not keep it. That fear is real, it is ancient, and it is worth taking seriously. But a fear held by everyone is not the same thing as an outcome measured in anyone. A proverb that appears everywhere describes the human condition, not your family's balance sheet.
The modern, quantified version of the curse traces overwhelmingly to one source: Roy Williams and Vic Preisser's 2003 book Preparing Heirs, which reported that 70 percent of wealth transfers "fail." That number then escaped the book and started reproducing in the wild, usually stripped of its definitions and its method.
James Grubman, a psychologist who has spent decades inside wealthy families and who has no interest in flattering them, went back and audited the claim. His 2022 paper in the International Family Offices Journal, There Is No 70% Rule, is the closest thing the field has to a formal retraction on the industry's behalf. His conclusion is worth quoting exactly: the 70 percent figure has "little empirical validity other than its presumed ubiquity."
Sit with that phrase. Presumed ubiquity. The number is believed because it is everywhere, and it is everywhere because it is believed. Grubman traces how the original research measured perceptions and definitions of failure that do not map onto actual wealth outcomes, how the sampling could not support a population-wide claim, and how decades of citation laundering turned a marketing statistic into received wisdom. Nobody tracked a representative cohort of wealthy families across three generations and counted the ruined ones. That study, the one everyone assumes exists, does not exist.
This matters commercially, not just academically. A large part of the wealth-services industry sells insurance against a disease whose epidemiology was never established. When your adviser opens with the 90 percent figure, ask for the cohort study. There is no shame in the question. There is considerable information in the silence that follows.
If the proverb were a law, the deep historical record should be a graveyard of elite surnames. Gregory Clark went looking for the graveyard and found something closer to a fortress.
Clark, an economic historian then at UC Davis, published The Son Also Rises with Princeton University Press in 2014. His method was clever precisely because it ignored what families say about themselves. He tracked rare surnames through centuries of records: probate registers, university enrollment, professional rosters, tax rolls. If elite status really burned off in three generations, elite surnames should appear in elite institutions at random population rates within a century or so.
They do not. Not remotely. Across England, Sweden, the United States, China, Japan, Chile, and India, Clark found that underlying social status persists with an intergenerational correlation of roughly 0.75 to 0.80. Conventional single-generation studies of income mobility had put the figure closer to 0.4, which made status look fluid. Clark showed that the one-generation number is noisy. Track the surname instead of the individual and the signal is brutal: families regress toward the mean so slowly that elite status takes 10 to 15 generations to wash out. That is 300 to 450 years, not 90.
His examples are hard to argue with. Descendants of Norman conquest surnames were still overrepresented at Oxford and Cambridge more than eight centuries after 1066. Swedish surnames tied to the 17th-century nobility remain overrepresented among Swedish physicians and attorneys today, in one of the most aggressively egalitarian societies ever engineered. Chinese elite surnames from the Qing era resurfaced at the top of the professional distribution within decades of the Cultural Revolution, an event specifically designed to destroy them.
Clark's work has its critics, and honesty requires naming them. Some economists argue that surname methods overstate persistence because they measure group averages rather than individual trajectories, and that rare-surname samples can carry selection quirks. Fair. But even the skeptical corrections leave persistence far above the folk number, and nobody in that literature defends anything resembling a three-generation wipeout of underlying status. The academic fight is over whether the fortress walls are 30 feet high or 50. Nobody thinks they are rubble.
Here is the reconciliation, and it is the most useful sentence in this piece. The proverb and the data are both right, about different things.
What decays fast is the liquid fortune. Money in its countable form gets divided among heirs, taxed at each transfer, spent on visible status, diluted by the simple arithmetic of reproduction. Two children become five grandchildren become thirteen great-grandchildren, and a fortune split thirteen ways while nobody adds to it looks, from the outside, exactly like a curse. The estate shrinks per capita even when it grows in total.
What decays even faster is cohesion. The founder's kitchen table, where everyone ate and argued and understood the enterprise, does not survive the founder. Cousins scatter across cities and countries. The shared story thins into anecdote. By the third generation the family is not a unit that owns things together; it is a list of people who share a dead ancestor and a lawyer. When cohesion goes, the assets follow, because nobody remaining has both the mandate and the trust to steward them jointly.
What does not decay, per Clark, is the substrate: educational expectations, professional networks, vocabulary, confidence, the unspoken knowledge of how institutions work and how money behaves. Grandchildren of wealth who "lose everything" still send their own children to good universities at wildly elevated rates. The bank account empties; the operating system persists. This is why the fourth generation of a "ruined" family so often rebuilds, and why Clark's surnames keep surfacing century after century.
The shirtsleeves proverb, read correctly, is not a prophecy about your family. It is an observation about the form wealth takes. Uncoordinated liquid wealth disperses. Embedded advantage endures. The question facing you is not whether your descendants will be fine, because on the long-run evidence they very probably will be. The question is whether they will be fine together, with the compounding power of shared assets and a shared name, or fine separately, as strangers who rebuilt alone.
Why does any of this matter beyond scorekeeping? Because families design differently under a curse than under a choice.
Families who believe the 90 percent number build defensively. They write trusts that read like restraining orders. They hide balance sheets from their own children until the reading of the will, on the theory that knowledge corrupts. They treat heirs as the threat model. And defensive architecture produces exactly the failure it fears: heirs who first meet the family wealth as adversaries, in a conference room, after the funeral, discover they have inherited assets but not judgment, structure but not story. The curse becomes self-executing. Grubman has made a version of this point for years: the fear of the proverb does more damage than the pattern it describes.
Families who believe the Clark number build differently. If advantage is durable and the fragile asset is cohesion, then cohesion is where the engineering budget goes. Regular family meetings with real agendas, not holiday dinners with an announcement at the end. Financial education that starts at twelve, not at probate. A written statement of what the wealth is for, argued over and revised, so the third generation inherits a purpose and not just a portfolio. Shared decisions made while the founder is alive to watch the second generation practice, fail small, and improve.
None of that is exotic. All of it is boring, repeatable, and cheap relative to the assets involved. It is also, on the evidence of the families who last, the actual difference. The proverb cannot be repealed, but it was never a law in the first place. It only ever described families who left cohesion to chance.
Stop planning around the curse. Retire the 70 percent line from your vocabulary, and when an adviser leads with it, ask for the study. Take the fear it encodes, because the fear is legitimate, and redirect it at the thing that actually dies in three generations: the family's habit of acting as one unit.
Then make it concrete. Put a date on the calendar within the next 90 days for a family meeting with an agenda, a financial topic your heirs have never been shown, and one decision made jointly. Cohesion is not a mood. It is a practice with a schedule.
This piece did its job if the next time someone tells you 90 percent of families lose it all by the third generation, you hear a proverb instead of a prophecy, and you go home and book the meeting anyway.