Retirees do not spend their savings. Not only the cautious ones or the underfunded ones, but the ones with plenty. Forty years of behavioural research keeps finding the same thing, and it has almost nothing to do with discipline. A dollar that arrives as income gets spent. A dollar that has to be sold first does not. The 4% rule is a rule about selling, which makes it a rule most people quietly decline to follow.
The People Who Would Not Spend
The retirement industry organises itself around the fear of running out. The data points the other way.
Banks, Blundell and Tanner documented the first half of it in 1998. Household consumption drops at retirement by more than the standard lifecycle model can account for, and it stays down. They called it the retirement savings puzzle, and the name stuck because nobody could explain it away.
The second half showed up in the asset data. Sudipto Banerjee, working from the Health and Retirement Study for the Employee Benefit Research Institute in 2018, tracked what happened to non-housing assets across the first twenty years of retirement.
Retirees who arrived with $500,000 or more had spent 11.8% of it at the median. The group with the least, under $200,000, spent the largest share and even that was around a quarter. Roughly a third of the whole sample finished the period holding more than they retired with.
These are people who saved successfully, retired on schedule, and then could not bring themselves to use the result. Whatever that is, it is not a funding problem.
Money Is Not Fungible
The lifecycle model the whole field rests on, from Modigliani and Brumberg in 1954, assumes a dollar is a dollar. Wealth is wealth whatever form it takes, and consumption gets smoothed against the total.
Shefrin and Thaler took that assumption apart in 1988. People do not hold one pot. They hold three, and they treat them as though they were different substances.
There is current income, which is what arrives this month. There is current assets, which is what has accumulated. There is future income, which has not arrived yet. The propensity to spend out of the first is close to one. Out of the second it is a few cents in the dollar. Out of the third it is near zero.
That is the finding, and it is larger than it looks. The same actuarial dollar is spendable or untouchable depending on which pocket it lands in. No amount of arithmetic moves it between them.
The Victorian Rule Was a Control Device
"Live off the income, never touch the capital" is older than modern finance, and it survived every attempt to argue it away.
Shefrin and Statman explained why in 1984, writing about why investors want cash dividends when selling a slice of the holding would do the same job. A retiree who spends dividends and refuses to sell shares is not confused about the arithmetic. They are running a rule that puts the principal out of reach, because a rule with exceptions in it is a rule that stops working the first time it is inconvenient.
The textbook calls this irrational. It is closer to a piece of self-control technology, and it does the job it was built for. What it costs is flexibility. What it buys is a decision that never has to be taken again.
The 4% Rule Fights the Instinct It Depends On
Bengen's rule is a rule about capital. Take four percent of the opening balance, raise it with inflation each year, repeat until the end.
Every one of those years requires a sale. Something has to be liquidated, on a day you pick, at a price you did not, and the balance is smaller afterwards. Nothing arrives. You take.
That is the framing loss aversion was named for. The withdrawal registers as depletion rather than payment, and it does so twelve times a year, for thirty years, against a number the retiree has spent their whole working life trying to increase.
It also fails hardest at the worst moment. In a drawdown the rule tells you to sell more units to raise the same real income, from a balance that has already fallen. That is precisely when the instinct to preserve the capital is loudest, and the rule has nothing to offer against it except an appeal to arithmetic that the retiree already understands and is ignoring anyway.
What a Ladder Does Differently
A ladder of inflation-linked bonds pays in a different shape. Coupons arrive on schedule. Principal arrives at maturity. The money lands in the account without anybody deciding anything.
Nothing is sold. There is no day to pick, no price to accept, no unit count going down. The balance falls as the ladder is consumed, and that fall is the design working rather than a warning about it.
The behavioural machinery that makes retirees hoard is disarmed by the instrument, at no cost, because the money genuinely is income. It arrives in the pocket with the high propensity to spend, and it does so for the same reason a pension does.
Banerjee's data shows the split inside a single sample. Across the first eighteen years of retirement, the median pensioner's non-housing assets fell 4%. For retirees without a pension the figure was 34%. The pensioners were not living smaller lives for it. They were funding the same retirement out of the pocket that opens easily, and leaving the one that does not alone.
Blanchett and Finke measured the size of the effect on spending directly, comparing households funding retirement from guaranteed income against households funding it from savings. The difference was not marginal. A dollar of income supported roughly twice the spending of a dollar left sitting in a portfolio.
The Same Dollar in Two Pockets
Take the 55-year-old from the first post in this series. She has $700,000 saved, plans to retire at 60 on $80,000 a year, and is priced against a 2.37% real yield as at July 2026.
At that rate a dollar per year of lifetime real income costs $15.83. Her Social Security entitlement is worth $267,000, which is $16,900 a year. Her portfolio is worth $700,000, which is $44,200 a year. Together they fund $61,100 against the $80,000 she planned.
Now watch what happens to those two numbers in practice. The $16,900 will be spent every year without a second thought, because it appears in a bank account on a schedule she does not control. The $44,200 will be agonised over, deferred, and quietly underspent.
Same woman, same year, same present value, priced off the same curve. The only difference is which pocket the money has to come out of, and that difference is worth more to her actual standard of living than most of the asset allocation decisions she will ever make.
The Illusion of Wealth Reverses
There is an obvious objection to all of this, and it is usually raised by quoting one half of a paper.
Goldstein, Hershfield and Benartzi ran the comparison in 2016. They showed people a sum of money either as a lump sum or as the monthly income it buys, and asked how adequate it seemed for retirement. At low wealth levels the lump sum won. A hundred thousand dollars reads as more adequate than $500 a month, which is the finding usually quoted as the illusion of wealth.
The second half is in the title of the paper and it goes missing from almost every summary. The pattern reverses. People are more sensitive to changes in wealth expressed as monthly income, so the two lines cross, and above the crossing point the monthly figure is the one that reads as adequate while the lump sum reads as thin.
In their between-subjects study that crossover sat at around $200,000, higher in the within-subject version, and its exact location moves with the rate at which income can be bought.
So the fear that income framing makes people feel poorer is a real finding about small balances. For a reader with $700,000 the paper points the other way. Their third experiment found intentions to save were lower at high wealth levels when the money was shown as income, which is the same thing as being readier to spend it.
The frame that gets the money spent and the frame that makes it look sufficient are, for this reader, the same frame.
Knowing You Are Funded Is Not the Same as Spending It
An earlier post in this series argued that Bill Perkins is right about dying with zero and short a denominator, and that the funded ratio supplies one. That still holds. It is also not sufficient.
Permission is not mechanism. Somebody told they are 118% funded has been given a fact, and that fact does not fill a bank account. On the first Tuesday of the month they still have to log in and sell something, and the whole weight of the research above says that is where the plan quietly stops.
The measurement answers whether the money is there. The form the money arrives in decides whether it gets used. Those are two different questions, and the retirement literature has spent most of its energy on the first.
What This Does Not Say
This names no product and recommends nothing. It is a finding about how people behave with money in different shapes, offered so that a reader who recognises the pattern in themselves knows what they are looking at.
It carries no implication that underspending is a failure either. Somebody who spends comfortably from a portfolio has nothing here to solve, and somebody whose surplus is deliberately earmarked for children or a foundation is not hoarding, they are funding a second liability that this measurement does not see.
The claim is narrower than any of that. If your plan assumes you will sell assets in retirement to fund your living, the evidence says that assumption is doing more work than the return assumption sitting next to it, and it has been tested far less.
A retirement plan has two halves. The first is whether the money is enough, which is a measurement, and this site exists to make it. The second is whether you will actually take it, which is not a measurement at all. It turns on whether the money arrives or has to be extracted, and the research is close to unanimous on which of those gets spent.
You can be fully funded and still live like you are not.
This is Post 14 of The Funded Ratio — a series on amifunded.com applying actuarial principles to individual retirement portfolios. Built on Shefrin and Thaler (1988) on the behavioural lifecycle hypothesis, Shefrin and Statman (1984) on the preference for cash dividends, Modigliani and Brumberg (1954) on lifecycle consumption, Banks, Blundell and Tanner (1998) on the retirement savings puzzle, Banerjee (2018) on asset decumulation through retirement, Goldstein, Hershfield and Benartzi (2016) on the illusion of wealth and its reversal, and Blanchett and Finke (2021) on spending from guaranteed income. No financial advice is given. All methods use publicly available academic research and market data.