How this calculator works

It answers one question: if you stop working and live off your savings, does the money outlast you? There is no single answer, because it depends on what markets do, and nobody knows that ahead of time. So instead of betting everything on one guessed average return, it replays your plan through every stretch of real financial history we have, from 1871 to today, and counts how often you would have come out fine.

Why replay history instead of picking an average

A retirement is a long string of good and bad years, and the order of those years matters as much as the average. Retiring just before a crash is a completely different experience from retiring just after one, even when the long-run average return is identical, because once you are living off the pot, a bad early run sells off shares you can never buy back. Averaging that away hides the exact risk you care about.

So the tool takes your plan (how much you have, how much you spend, how it is invested) and runs it as if you had retired in 1871. Then again as if you had retired in 1872, then 1873, and so on through every starting year for which there is enough history left to cover your whole retirement. Each of those runs is one cycle. With roughly 150 years of data and a 40-year plan, that is about 110 overlapping cycles, 110 complete retirements that actually could have happened.

What happens in a single year

Inside every cycle the calculator walks one year at a time. The order is deliberate and matches the way the reference tool (engaging-data's "Rich, Broke or Dead") does it, so the results line up:

Money comes out before the year's growth is applied, which is the cautious assumption: you cannot spend a return you have not earned yet. Cash earns a real short-term interest rate (not zero, and not magically tracking inflation), and fund fees are charged on stocks and bonds only, never on cash. A cycle "fails" the first year the balance hits zero.

Reading the chart: rich, broke or dead

The main chart stacks up, for each age, what fraction of those ~110 cycles ended in each state. Read a vertical slice at any age as "out of everyone who retired with this plan, here is how they were doing at this age":

The greens are success, measured against your starting balance in today's money: a thin light-green band means the money survived but shrank, a thick dark-green top means it grew several times over. The red band at the bottom is the share that went broke. The grey band is different: it has nothing to do with money. It is the chance, from population mortality tables, that you are simply no longer alive at that age. It matters because "the plan fails at age 97" is not much of a failure if most people are gone by then. So the mortality curve is drawn straight onto the chart as the grey band: you can weigh any failure against the odds you were still alive to see it, and a plan that only breaks deep in old age is barely a failure at all.

What the money is invested in

You pick a stock index (the US S&P 500, or MSCI World), a bond type (US Treasuries, or a developed-world government bond index), and how much sits in cash. Every series is a real total return: dividends and coupons reinvested, then adjusted for that year's inflation, so a "7%" year means 7% of actual purchasing power gained. The full year-by-year data, and a growth-of-1 chart, live on the Datasets page.

Currencies need one honest caveat. There is no clean century of euro or franc stock-market data, and the historical guilder/dollar and franc/dollar exchange rates are mangled by the gold standard and the world wars (the guilder more than doubled against the dollar in a single post-war year). Splicing those in would inject volatility that never really hit a long-term investor. So euro and franc results use the same real asset returns as the dollar, then re-apply each country's own inflation and its own local cash rate. Over long horizons real exchange rates barely drift, so this is closer to an investor's lived experience than pretending the broken historical FX was real. The practical effect: switching currency mostly changes the inflation backdrop and the cash rate, not the underlying market.

Historical or Monte Carlo

By default the calculator replays the actual sequences of history, one cycle per starting year. The Monte Carlo option does something related but different: it stitches together a thousand new sequences by drawing random ten-year blocks of real history and gluing them end to end. It still uses real returns, and keeps real crashes intact within each block, but it explores many more orderings than the handful history happened to deal. Historical answers "how would this plan have fared in the runs that actually happened"; Monte Carlo answers "how does it fare across many plausible reshuffles of the same data". Looking at both is healthy: history can be too kind (it is really only a few independent 40-year runs), while a reshuffle can be a touch too wild (it breaks the gentle mean reversion real markets show). Neither is the truth; the answer usually sits between them. You can also confine the reshuffle to a more recent window, the modern S&P era from 1928 or the live-MSCI-World era from 1970, for when the thin deep-history proxy feels like too much of a stretch.

Taxes, the part most calculators skip

This is the feature the tool was built for. The same portfolio can land very differently depending on how it is taxed, so all the regimes are computed at once and shown side by side; click any row in the table to make it the active one. They differ mainly in what they tax and when:

• Capital gains. Only the gain you actually realise when you sell is taxed, and only the gain portion, not the part that is just your own money coming back. Everything you have not sold keeps compounding untaxed. Because unsold growth is never touched, it is among the lightest here for a long holder.
• Dutch Box 3 (2028+). The proposed "real return" wealth tax charges you every year on your whole nominal return, including gains you have not sold. Because it taxes nominal (not inflation-adjusted) returns, inflation alone can create a taxable "gain", which is why it bites hardest over long horizons.
• Beleggings-BV. Holding investments inside a Dutch company. Corporate tax hits the gain when it is realised, then a second dividend tax hits what you pay out to yourself. Two layers, but again only on gains and profit, and unrealised growth defers until you sell, which is its edge over Box 3.
• Swiss. Private capital gains are not taxed at all. Instead there is an annual wealth tax, income tax on the portfolio's dividend yield, and mandatory social-security contributions before 65. Modelled on Zürich by default, with presets for nine other cantons and a single/couple toggle; the franc thresholds are converted to whatever currency you pick.
• Germany. A flat 26.375% (Abgeltungsteuer plus the solidarity surcharge) on realised gains, softened by a 30% tax-free slice for equity funds and a small annual allowance. The twist is the Vorabpauschale: each year a slice of the fund's notional return is taxed up front even though nothing was sold, then credited back when you finally do sell. It is a gentler cousin of Box 3, only the notional yield is touched, and only when the fund actually rose. Germany also has no AHV-style pension levy on a Privatier, but statutory health and care insurance is mandatory and, for a voluntary member, charged on capital income (~21%, floored and capped), a separate cost on top of the tax that often outweighs it.
• Belgium (2026 reform). A new 10% tax on realised gains on financial assets, but only on the slice above a €10,000-per-person yearly exemption (doubled for couples), so a long equity holder pays little. The catch is interest: cash interest and the coupon part of bond funds are taxed separately at 30%, so Belgium treats shares far more kindly than cash or bonds, and there is no wealth tax.

Every monetary threshold in these rules (exemptions, brackets, allowances) grows with inflation in the simulation, the way real tax systems adjust over time, so a multi-decade run is not quietly dragged into higher brackets by inflation alone.

The numbers under the table

• Money lasts is the share of cycles where the balance never hit zero.
• Median left is the middle cycle's ending balance in today's money, with the 25th-to-75th-percentile range (the middle half of outcomes) beneath it, because the average alone hides how wide the spread is.
• Tax paid is the median lifetime tax across cycles, also in today's money.
• Yearly drag expresses that tax as an equivalent annual fee on the portfolio. It is not a flat percentage, because the tax rules are progressive: spending more shrinks the pot into lower brackets, spending less lets it compound into higher ones, so the drag moves with the path.
• Safe spend. Flip the table's Compare control to "max safe spending" and each row shows the most you could spend per year and still last 90% of the time, with the other columns recomputed at that spend. Click a row to drop that amount into your plan.

What it can and cannot tell you

This is a model, not a forecast or financial advice. History is the best evidence we have, but the future is not obliged to rhyme with it: the next 40 years could be worse than anything in the record. The cycles overlap and share years, so they are not fully independent samples. Tax law, especially the not-yet-enacted Dutch regime, can change. Mortality is a population average, not you. And it models a steady mechanical plan, not the real human tendency to adjust, panic, or carry on working. Treat the output as a way to compare choices and see the shape of the risk, not as a promise.

Where the data comes from

• Long-run market returns and inflation: the Jordà-Schularick-Taylor Macrohistory Database (1870–2020), the US series cross-checked against Robert Shiller's data, and extended to 2025 with FRED, MSCI, FTSE, ECB, SNB and Eurostat figures. Please cite Jordà, Schularick and Taylor (2017), "Macrofinancial History and the New Business Cycle Facts", and for return data Jordà, Knoll, Kuvshinov, Schularick and Taylor (2019), "The Rate of Return on Everything, 1870–2015", QJE 134(3); US market and inflation data from Robert Shiller, Yale University.
• Mortality: Eurostat life tables for Europe and the US Social Security tables, with healthy and smoker variants scaled from real per-age ratios.
• Taxes: the Dutch Wet werkelijk rendement Box 3 proposal, Belastingdienst and PwC figures for Vpb and Box 2, and the Zürich tax law, ESTV cantonal data and the official AHV contribution table for Switzerland.
• The whole approach follows engaging-data's "Rich, Broke or Dead", rebuilt from primary sources and extended with tax regimes and currencies. The US, S&P, no-cash case reproduces it to the decimal.

All charts on this page are generated live from the calculator's own engine and data.