Die With Zero Math, Priced in Unspendable Dollars

The 4% rule is an insurance contract written against the worst year in modern market history, and like most insurance it is priced for the catastrophe while the median policyholder overpays. The premium is collected in a strange currency: working years nobody needed, and purchases whose windows closed before the budget allowed them. Die with zero usually arrives as a pep talk about bravery and bucket lists. The version that should move a numerate saver is computational, because the numbers are more confrontational than the slogans. Across historical simulations, the median retiree who withdrew an inflation-adjusted 4% finished with more real wealth than they retired with, often a multiple of the starting portfolio, while the capacity to spend it was expiring on a schedule nobody printed for them.

Three questions follow, and all three have arithmetic answers rather than attitudinal ones. How many unspendable dollars does a worst-case-calibrated plan historically leave behind? How many work years does that surplus represent at your savings rate? And which purchases should move forward on the calendar before age closes them regardless of net worth? The honest objection gets its own section: part of the cushion is rational insurance for a 50-year retirement, and the goal is to price that insurance, not delete it.

The Worst-Case Math Behind the 4% Rule

William Bengen's 1994 analysis asked a survival question: what initial withdrawal rate, then raised with inflation every year, would have survived every 30-year retirement in modern US market history? Roughly 4%, for a balanced portfolio, and the Trinity study later confirmed the pattern across asset mixes. Understand what that means. The rule was reverse-engineered from the worst cohorts, the retirements that began into the crash of 1929 or the stagflation of the mid-1960s. Surviving those sequences is the entire design specification. Nobody optimized for the median.

Calibrate to survive the worst and you overfund everyone else. The simulations behind the rule do not squeak through at the median; they finish rich. Backtests of inflation-adjusted 4% withdrawals over 30-year horizons commonly show median terminal wealth on the order of twice the starting balance in real terms, sometimes more. The safe withdrawal rate series at Early Retirement Now walks through this cohort by cohort. So how much money do retirees die with under this plan? In simulation, the median answer is most of it, and then some.

Three mechanisms stack to produce that result. Median market returns beat the disastrous path the rule was braced against. Withdrawals held flat in real terms outrun what retirees actually spend as they age, which the next section quantifies. And the catastrophe the rule carried insurance against simply never arrived in most cohorts. The surplus is the unclaimed payout on a policy whose disaster, in the median timeline, did not occur.

The Retirement Spending Smile, in Real Dollars

Does retirement spending decrease with age? For most households, the consumption data says yes, most of the time. David Blanchett's research on the retirement consumption puzzle found real household spending falls through mid-retirement on the order of one to two percent a year, a cumulative real decline of roughly 20 to 25 percent by the mid-80s before health costs bend the curve back up. Plotted by age, the line smiles. It falls through the go-go years, flattens through the slow-go years, and ticks upward with medical spending in the no-go years. Practitioners call it the retirement spending smile.

The causes are not mysterious. Commuting and work costs vanish, the mortgage amortizes to zero, and the children become self-funding. Meanwhile the spending that is genuinely age-sensitive, meaning travel, adventure, and physically demanding hobbies, concentrates early in retirement while health and companionship permit it, then thins. Official expenditure surveys show the step-down plainly: average household spending in the 75-and-older band runs well below the 65-to-74 band, with the cuts concentrated in transportation, travel, apparel, and entertainment, partly offset by higher health outlays.

The planning consequence is blunt. A 25x-expenses target assumes you will spend this year's budget, inflation-adjusted, at age 85. The evidence says most households will not. Funding a budget that declines about one percent a year in real terms is meaningfully cheaper than funding one held flat forever, which means a smile-adjusted plan defends a higher initial withdrawal rate than a flat one at the same historical safety. The finish line you are grinding toward is, for most people, overstated by construction.

Your Unspendable Number, in Dollars and Work Years

The cost of oversaving for retirement is usually described as anxiety. It should be described in dollars and in years, and both are computable. Work everything in real terms, today's money, in four moves.

1. Estimate median terminal wealth. From historical simulations at your planned withdrawal rate, what did the median cohort finish with? At 4% over 30 years, a defensible working assumption is on the order of 2x the starting balance in real terms. Scale down for 50-year horizons and conservative allocations.
2. Estimate smile-adjusted lifetime spending. Replace the flat budget with one declining about 1% a year in real terms into your early 80s, then flattening with a late health bump.
3. Subtract. What the median path produced minus what the declining path consumed is your unspendable number.
4. Divide by annual savings. That converts abstract cushion into work years, the price of the insurance in the only currency you cannot earn back.

Unspendable dollars = median terminal wealth minus smile-adjusted lifetime spendingWork years = unspendable dollars divided by annual savings

This is not purely a simulation artifact. EBRI's decumulation research followed households deep into retirement and found a substantial share held as much or more in assets than on the day they retired, despite decades of drawing income. Dying with money you never spent is the observed median behavior of cautious retirees, not a theoretical risk.

Two worked examples, with the assumptions stated so you can substitute your own.

Example one. A saver targets a $1.2M portfolio supporting $48k a year, currently saving $50k annually. If the median path at 4% ends near twice the start in real terms, terminal wealth is about $2.4M and the plan created roughly $1.2M more real wealth than the retirement consumed. Funding that gap took 24 years of $50k contributions, or about 17 calendar years once 5% real returns do part of the work. That is how many extra working years the worst-case insurance cost, if your path had been the median one.

Example two. A higher earner targets $2M supporting $80k, saving $100k annually. Same assumption: about $4M terminal wealth, about $2M of surplus, 20 contribution-years, or roughly 14 calendar years at a 5% real return.

These are illustrations, not forecasts. The honest range around any median is wide, and the multiple can be smaller for long horizons and conservative mixes. Run your own allocation and spending path through a historical simulator, and use a conservative multiple if you want a floor instead of a point estimate. The exact figure matters less than the discipline of computing one at all.

The Capacity Calendar, Decade by Decade

Spending windows that close with age are the second half of the problem, and that half is not denominated in dollars. Spending capacity is age-gated. After roughly 75, both discretionary outlays and physical capability tend to fall steeply, and categories like adventure travel have windows that close regardless of wealth. Waiting until you can comfortably afford something is often indistinguishable from waiting until it is impossible.

One line bends the wrong way. The late-life spike in health and custodial spending, including long-term care, is the one window that opens instead of closing, which is why EBRI's work on retiree views on long-term care belongs in your stack even in a spend-it-down article. It is the one category you cannot pre-spend.

The table's use is personal, not decorative. List your five largest age-gated expenditures, put a closing date on each, and you have converted a vague fear of dying with money into a set of deadlines. Deadlines get met. Vague fears get deferred.

Why the Surplus Is Partly Rational Insurance

Now the steelman, because the arithmetic above is dangerously persuasive in isolation.

First, horizons. The 4% figure was calibrated on 30-year retirements. Someone retiring at 40 faces 50 or 60 years, and the same simulation series cited earlier finds historically sustainable rates fall materially below 4%, toward the mid-3s, as horizons stretch, depending on allocation and starting valuations. An early retiree spending 3.5% is not being timid.

Second, and more subtly, the median surplus is not a defect of the rule. It is the mechanism. A rule calibrated to end the median cohort at zero would have failed the worst cohorts outright. The wealth left over in lucky timelines is the same wealth that rescued the unlucky ones. Sequence-of-returns risk means the first decade of returns dominates outcomes, and a retiree who hits a 1929 or a 1968 in year one spends from a shrinking base for decades. The cushion you die with on good paths is the payout that saved the bad ones.

Third, structure. Most FIRE households run a single portfolio with no pension, thin annuitized income, and Social Security a distant 25 years off. Returning to work in your 60s after a decade away is a one-way door that rarely opens onto the old salary. Part of the surplus is simply the price of walking through retirement without a backstop.

The actionable output is a split, not a slogan. Carve the plan into a floor you treat with full worst-case discipline, and a discretionary layer you deliberately spend against the calendar while the windows are open. Price the insurance consciously instead of letting it silently claim your entire surplus.

Die With Zero Deliberately, With a Calendar

Bill Perkins' Die with Zero argues that money's value is time-stamped: experiences pay a memory dividend, giving while living beats giving at death, and accumulating beyond your capacity to convert dollars into lived life is a quiet error. As philosophy it is fine. Most readers lack the machinery, which is four moves.

1. Date your windows. Take the five biggest age-gated items on your list and write the age after which each becomes impossible or joyless. A Kilimanjaro trek has a different deadline than season tickets.
2. Pre-fund the closing windows now. During accumulation, run sinking funds for the nearest-closing categories. Pull the trip forward from 67 to 52. The spreadsheet forgives you before your knees do.
3. Recut the plan on a declining real budget. Build retirement spending as a smile-adjusted path rather than a flat inflation-adjusted line. Lifetime cost falls, the defensible initial withdrawal rises, and the finish line drops below the flat 25x you were grinding toward.
4. Recompute the unspendable number every year or two. Each year of actual returns narrows the distribution of outcomes. A strong first decade releases spending that a weak one locks down. That is guardrails, not vibes.

Then close the loop the opening paragraphs set up. Recompute the number annually and set it beside the capacity calendar. If unspendable dollars keep growing while the open windows keep shrinking, the plan is funding an heir's windfall with years you could have lived. That trade is sometimes correct. It should never be unexamined.