Will my money last? A practical guide to stress-testing your retirement plan.
Ask retirees what keeps them up at night, and "running out of money" beats nearly everything else. It's a reasonable fear: retirement is the one financial project you can't redo.
The good news is that "will my money last?" is an answerable question — not with certainty, but with rigor. A plan stress-tested three different ways is worth more than one built on a single smooth projection. Here's how to do it.
Why isn't a single projection enough?
Most retirement calculators show you one line: your balance gliding smoothly upward at 7% a year while you withdraw a fixed amount. It always works. It's also fiction.
Real markets don't deliver averages; they deliver sequences — crashes, recoveries, inflation spikes, long flat stretches, rate-hiking cycles that reprice your bonds overnight. Two retirees with identical average returns can end up nearly $300,000 apart over a decade, purely because of the order those returns arrived in. A single projection assumes the order away.
A serious stress test asks three different questions instead of one:
- What if I define the worst case myself? (Custom crash test)
- What would have happened to someone like me in every year of recorded history? (Historical backtest)
- What could happen across thousands of plausible futures, including ones we've never seen? (Monte Carlo)
Each test has blind spots. Together, they cover for each other.
Test 1: Define your own crash
Generic bear-market assumptions — "assume a 20% drop" — are better than nothing, but they're abstract. Your fear isn't abstract. It's specific: What if the market crashes 30% the year after I retire and takes four years to recover? What if inflation runs hot at the same time, like the 1970s?
A useful crash test lets you specify the downturn yourself: how far, how fast, how long the recovery takes, and whether inflation spikes alongside it. Then it runs your plan — your accounts, your withdrawal order, your Social Security timing — through that scenario, year by year.
What you're looking for isn't just "did I survive." It's where the damage happens:
- The early years are the danger zone. A crash in year two of retirement does far more damage than the same crash in year fifteen, because early withdrawals compound against early losses.
- Withdrawal order matters. Selling stocks in a crash to fund spending locks in losses. Plans that pull from cash or bonds first — or that pause Roth conversions during the dip and resume after — weather the same crash better.
- Inflation is the silent killer. A 25% market drop with 2% inflation is recoverable. The same drop with 8% inflation forces your withdrawals to grow every year while the portfolio shrinks. Always test the combination, not just the crash.
If your plan survives your own worst-case scenario with margin to spare, you can spend with more confidence. If it breaks, you now know exactly which lever to pull — and you can pull it before the crash, not during.
Test 2: Replay history
The historical backtest is the oldest stress test in retirement planning, and still one of the most persuasive. The idea: take your plan and run it as if you had retired in every starting year on record — 1872, 1873, 1874, all the way to 2025 — using the actual market returns and inflation that followed each year.
Runway's Historical Backtest does exactly this across 125 starting years, built on the same long-run U.S. market dataset (monthly stock prices, dividends, earnings, and inflation back to 1871) that economist Robert Shiller maintains at Yale and that most retirement research of this kind is built on. The output is a distribution of outcomes, not a single answer. You'll see the retiree who started in 1982 — the beginning of the greatest bull market in history — and the one who started in 1966, who endured 15 years of flat markets and 1970s inflation. You'll see where your plan lands in that spread: does it survive the 1966s, or only the average years?
What history teaches you:
- The worst case is worse than you think, and survivable anyway. The 1966 retiree needed a ~4% withdrawal rate to make it 30 years. That's the number the entire 4% rule is built on. If your plan survives the 1966–1995 window, it's survived the worst sequence in 150 years of U.S. data.
- Most starting years were fine. The median historical outcome for a 4% withdrawer was dying with nearly triple the starting balance. History's lesson isn't just "be careful" — it's also "don't be so careful you don't live."
- Inflation regimes matter as much as market crashes. The backtest's worst periods pair bad markets with high inflation. A test that only randomizes stock returns misses half the story.
History's limits deserve honesty too. One hundred twenty-five starting years is one country, one century and a half, with heavy overlap between adjacent 30-year windows. The future could produce a sequence worse than 1966 — Japan's post-1989 experience is a standing reminder that "stocks always recover within a decade" is a U.S.-specific observation, not a law of nature. The backtest tells you what did happen, not the full range of what can happen. That's what the third test is for.
Test 3: Monte Carlo — thousands of futures, including ones we've never seen
Monte Carlo simulation builds thousands of hypothetical return sequences from statistical assumptions about average returns, volatility, and inflation — then runs your plan through all of them. Runway's Pro tier runs 2,000 simulations; the free tier runs 200.
Where the backtest replays what happened, Monte Carlo explores what could happen: sequences worse than 1966, crashes that arrive in year one instead of year three, inflation spikes paired with flat markets in combinations history never dealt. If the backtest is a highlight reel, Monte Carlo is the full possibility space.
What Monte Carlo captures that history can't:
- Tail risk. History gives you one sample of reality. Monte Carlo generates the 1-in-100 sequences — where three bad things happen at once.
- Your specific plan's fragility. A plan with rigid inflation-adjusted withdrawals fails differently than one with flexible spending. Monte Carlo shows you the failure pattern: does the plan break in 5% of scenarios, and are those failures clustered in early-retirement crashes?
What Monte Carlo misses — and you should know this before trusting it:
- It's only as good as its assumptions. If you assume 7% average returns with 15% volatility, the simulation obediently produces futures consistent with those inputs. The assumptions deserve as much scrutiny as the results.
- It typically assumes returns are random each year. Real markets have regimes — long stretches where valuations, yields, and inflation behave differently. Most engines don't model regime shifts well.
- Fat tails are hard. Extreme events happen more often than bell-curve statistics predict. A naive engine understates the worst cases.
The right way to use Monte Carlo is as a complement to history, not a replacement: history grounds you in what actually happened, Monte Carlo probes beyond it.
Historical vs. Monte Carlo: use both, trust neither blindly
| Historical backtest | Monte Carlo | |
|---|---|---|
| Asks | What happened to retirees like me in every recorded year? | What could happen across thousands of plausible futures? |
| Strength | Real returns, real inflation, real correlations — no assumptions | Explores sequences worse than anything on record |
| Weakness | One country, one history; can't show what hasn't happened | Only as good as its return assumptions; weak on regimes |
| Best for | Calibrating against the worst known sequence (1966) | Finding your plan's breaking points and tail risk |
If your plan survives 125 historical starting years and shows a strong success rate across 2,000 simulated futures, you've done more diligence than the vast majority of retirees. If the two tests disagree — history says fine, Monte Carlo says fragile, or vice versa — that disagreement itself is information: it tells you where your plan's assumptions are doing the heavy lifting.
What success rate should you aim for?
This is where honest nuance matters, because there's no magic number.
Morningstar's retirement research uses a 90% probability of not outliving your money as its benchmark. Bengen's original work effectively demanded 100% on the historical record — survival in every tested period. Many financial planners treat 80–90% as the comfort zone. The tradeoff:
- Below ~70%: fragile. A below-average sequence likely breaks it. Action warranted.
- 80–90%: robust to most futures. A reasonable place to be.
- Above ~95%: ask what it's costing you. A 99% success rate usually means dying with far more than you started with — working longer than needed or spending less than you could have. Excessive caution has its own failure mode: regret.
There's also a subtlety in what "success" means. Most calculators define it as "balance stays above zero." But a plan that survives with $1,000 left and one that survives with $1 million left both count as "success." Look at the distribution of ending balances, not just the pass rate. If your median outcome is triple your starting balance, the binding question isn't survival — it's whether you're spending enough.
And a success rate is a model output, not a guarantee. Treat the number as a compass, not a contract.
From results to action: five concrete moves
A stress test that doesn't change your behavior is entertainment. Here's what to do with the results:
1. Write down your response rules in advance. If the backtest shows your plan breaks when markets fall 25% in the first three years, decide now what you'd cut: which discretionary spending goes first, whether you'd pause Roth conversions, whether you'd pick up part-time income. Pre-commitment beats panic.
2. Attack the tax drag before RMDs begin. Under SECURE 2.0, traditional IRA and 401(k) owners born 1951–1959 face required minimum distributions at 73; those born 1960 or later at 75. Those forced withdrawals are taxed as ordinary income whether you need the money or not. Systematic Roth conversions in your 60s, guided by a bracket-by-bracket comparison, shrink the future RMD problem. On a typical $1.4M plan, Runway's Roth conversion optimizer has found around $180,000 in lifetime tax savings — or get a quick first read with our free Roth conversion calculator before diving into the full bracket-by-bracket search.
3. Time Social Security deliberately. Every year you delay Social Security past 62 (up to 70) increases your inflation-protected lifetime income — income that doesn't depend on markets at all. Runway's full claiming-age grid shows the tradeoff year by year. For sequence risk, a larger guaranteed benefit means a smaller portfolio withdrawal in exactly the years when withdrawals are most dangerous.
4. Stress the non-market risks too. Markets aren't the only threat. Runway's long-term-care stress scenario models what an extended care event does to the surviving spouse's plan — often the true worst case, and one that pure market testing misses. If you're 60–65, ACA premium tax credit modeling matters too: managing your income in those bridge years can be worth thousands in subsidies.
5. Re-test annually. A stress test is a snapshot, not a tattoo. Markets move, spending changes, tax law changes. Re-running your crash tests and backtests once a year — or after any big life change — keeps the plan honest. Bengen himself recommends reviewing withdrawals against your plan's template every couple of years at minimum.
How to run all three tests on your own plan
This is exactly what Runway was built for. The custom Market Crash Test lets you define your own downturn and watch your plan absorb it year by year. The Historical Backtest runs your plan through 125 real starting years from 1872 to 2025. The Monte Carlo engine runs 2,000 simulated futures on Pro (200 on the free tier). And multi-scenario comparison lines up "spend $80k" against "spend $95k" side by side, so the tradeoffs are visible instead of abstract.
Start free — no card required — with the 200-simulation Monte Carlo, the 30-year projection table, and withdrawal sequencing. Define your own crash. See what survives. Then decide what to do about the rest.
Stress-test your own plan. Run a custom market crash, a 125-year historical backtest, and a Monte Carlo simulation against your real accounts — free to start, no credit card required.
Try the planner