Family Spending Rule After an Exit: Bengen's 30-Year Math

A family spending rule after an exit cannot be borrowed from retirement planning. The familiar 4% figure came out of a study asking whether one retired household's portfolio could survive roughly 30 years without running out.¹ Capital meant to reach grandchildren needs its own rule — one that names a longer horizon, lets the annual draw fall when the portfolio does, and counts how many households will eventually be drawing on it.

In 1994, a financial planner named William Bengen published a paper in the Journal of Financial Planning² with a deliberately narrow question. Using American stock and bond records going back to 1926,³ he asked: if a retiree withdrew a fixed percentage of the starting portfolio in the first year, then raised that dollar amount with inflation every year after, what was the highest rate that would have lasted at least 30 years no matter which year the retirement began?⁴ For a portfolio split between large-company stocks and intermediate-term government bonds, the answer was about 4%.¹ The finding was honest and tied tightly to its assumptions. Then it escaped. Within a decade it was "the 4% rule," and it is now quoted in family offices by people whose capital is meant to outlive them by a century.

There are dozens of questions a family could ask about spending. Three of them decide whether Bengen's math applies at all, and all three come before anyone names a rate.

How long is this capital actually being asked to last?

The governing principle: the horizon is the first input, not the rate. Thirty years was not a number Bengen discovered; it was one he chose, because a couple retiring at 65 rarely needs a portfolio to keep working past 95.⁵ And the standard he tested was survival — not running out inside the window. The test tolerated depletion; it did not aim for it. A portfolio that still held a dollar at the end passed, whether that dollar was the last one or the first of a large remainder.

That is the right definition of success for a retiree. It is a definition of catastrophe for a family that intends the capital to fund people who are not yet born. And the arithmetic moves against longer horizons: the further out the finish line, the lower the draw that history would have supported, because a bad early decade has more years in which to do its damage and no scheduled end date to rescue the account. A family drawing at a retiree's rate is running a 30-year plan against a hundred-year obligation, and nothing in the original research says that works.

The concrete version. A founder sells in her late fifties. She has two adult children and four grandchildren under ten. The youngest of those grandchildren could plausibly still be drawing on this capital in the 2090s. Bengen's retiree had one lifetime to fund; this family has three, overlapping, with the last one barely begun. The real question is not "what rate" but "whose lifetime ends the test" — and for most families in this position, the honest answer is nobody's.

What changes if the answer is 30 years? If the family genuinely intends to consume the capital within the founders' lifetimes, with no legacy beyond gifts already made, the retirement literature is the right shelf and Bengen's framing is roughly the right instrument. That is a legitimate choice. It simply needs to be made deliberately and written down, so that heirs do not inherit an impression of permanence the math was never asked to support.

Can the annual check fall when the portfolio does?

The governing principle: a rule that ignores the portfolio's present value cannot correct itself. In Bengen's design, the retiree fixes a dollar amount in year one and never looks back. It rises with inflation whether the portfolio has doubled or halved. That rigidity was the point — a retiree wants a steady paycheck, and the study was searching for the largest steady paycheck history had not broken. What breaks such a plan is rarely the long-run average; it is sequence, the order in which good and bad years arrive once money is leaving.

Institutions that expect to exist indefinitely handle this differently. University endowments typically spend a target percentage of market value averaged over several trailing years, often blended with the prior year's budget so the number cannot lurch. The structure matters more than the percentage. When the capital falls, the draw falls with it, after a lag. When it rises, the draw rises, after a lag. The institution is never surprised into paying a fixed check out of a shrunken base.

Founders already understand this, because it is how they ran their companies. Owner distributions came out of what the business actually produced that year, not a figure agreed at the founding. Nobody paid themselves a fixed, inflation-linked dividend through a bad year. One other figure gets borrowed in the same careless way: private foundations must generally distribute at least 5% of their investment assets each year for charitable purposes.⁶ That is a legal floor for a charitable entity, not evidence that a taxable family fortune can spend 5% and hold its purchasing power.

The year of an exit is the worst possible moment to freeze a dollar amount. Proceeds have just become cash, lifestyle tends to expand to meet liquidity, and a first-year number raised with inflation forever turns a celebration into a permanent payroll.

What changes if the answer is no? Some families carry genuinely fixed obligations — a school commitment, a property with unavoidable upkeep, a relative whose support is not negotiable. Where spending truly cannot flex, that portion behaves like a fixed charge the capital must cover, and the family needs to know in advance how much of the total is pinned by it. The retiree's rigid math belongs to that pinned portion only, never to the whole. And if the check can fall, the rule should say who holds the authority to cut it, against what measure, and how that decision gets explained to the people who feel it.

How many households will be drawing on this, and what counts as a draw?

The governing principle: a spending rule names its beneficiaries and counts every transfer, not just its percentage. Bengen's retiree was a single household, and the unstated assumption was that the number of claimants would only ever decline. Families run the other way. Two children become two households. Four grandchildren become four more. A fortune that supported one dinner table may, by the third generation, be asked to support a dozen, each with its own view of what a fair draw looks like.

The example is arithmetic any founder can do in their head. Assume two children per household. A family whose rule reads "a percentage of the total, divided among whoever is alive" has quietly written a rule in which each household's share halves every generation, before a single market cycle is counted. A family whose rule is set per household has quietly written one in which the total draw doubles every generation. Neither is wrong. But a family that has not noticed the difference has not actually written a rule. As the Pritzker story shows, headcount is ultimately a governance question wearing a spreadsheet.

Counting the claimants is only half of it; the claims have to be counted too. Money that leaves as a loan to a child and is never collected is spending, even though it never appears in the annual figure the family believes it is obeying — the loan nobody writes down. Nor is the famous study's setting the family's setting: it assumed tax-deferred retirement accounts, while sale proceeds usually sit in taxable ones, where funding the check can mean selling an asset and paying tax on the gain. That tax is consumption. A rule that ignores either leak has a hole in it.

What changes if the answer is "just us"? A founder with no children, or whose heirs are provided for by other means, sits much closer to Bengen's retiree than the rest of this piece suggests. The horizon shortens, the claimants stay fixed, and it becomes a lifetime question again. For everyone else, the beneficiaries multiply, and the rule has to say what happens when they do — well before the second generation arrives to ask.

What to do with the three answers

Write them down before anyone proposes a number. The horizon decides whether the retirement literature applies at all. The fixed-or-flexing choice decides whether the rule can survive a bad decade. The headcount, and an honest tally of what counts as a draw, decides what "the family" will mean by the time the rule matters most. A family that can write one clear sentence on each has something an advisor, a trustee or a child can be held to. Only then does a percentage belong in the conversation — and by then it will be obvious that 4% was the answer to somebody else's question. Bengen never claimed otherwise. The families who borrowed it did.