How Much Do You Actually Need to Retire?
Most people size this number backwards — pick a target, then try to hit it. Working from your actual spending outward gets you a better answer, and a more useful one.
- There's no single "retirement number." The number depends on your spending, your retirement age, your Social Security timing, and how much confidence you want in the outcome — change any one and the target moves.
- Most calculators start with a target ("do I have $1.5M?") and work backward. The more useful question runs the other direction: given what I've actually saved, what monthly spending does my plan support?
- Spending itself is usually the weakest input in the whole exercise — people guess it from memory, and it often isn't flat across retirement the way most tools assume.
- You can run the inverse calculation on your own numbers below — free, no signup, nothing leaves your browser.
You're probably sizing your number backwards
Ask most people how much they need to retire and they'll answer with a target: "$1.5 million," "$2 million," some round number picked up from an article, a coworker, or a rule of thumb. Then the plan becomes: save toward that number, and hope it's the right one.
That's backwards. The target came first and the actual retirement — what it costs to live the life you want — got fit around it after the fact, or not examined closely at all. A number chosen before you know your real spending isn't a target, it's a guess wearing a target's clothes.
The more useful order runs the other direction. Start with what your life in retirement actually costs — real spending, not a memory of it. Add in what you've saved, what you're still contributing, when you plan to stop working, and when you'll claim Social Security. Only then does a number worth trusting fall out the other end. It's not a different kind of math. It's the same math, run in the order that actually answers the question you're asking.
Why spending is the hardest input to get right
Every retirement calculation is only as good as its spending number, and spending is the input people are worst at estimating — not because they're careless, but because almost nobody tracks it closely enough to know it cold. Most people answer "what do you spend a month?" from memory, rounded to a number that feels roughly right. Memory is a bad instrument for this. It smooths out the irregular months, forgets the annual costs that don't show up monthly (property tax, insurance premiums, the roof), and anchors on whatever number was true a few years ago.
The fix isn't complicated, just unglamorous: pull three to six months of real bank and card statements and total what actually left the account, including the annual expenses divided down to a monthly figure. That number is almost always different from the guess — sometimes lower, often higher, but rarely the same.
Spending isn't flat, either
The second mistake compounds the first: most calculators, and most people's mental model, assume retirement spending is one flat, inflation-adjusted number held constant for 30 years. Real retirements don't look like that. A widely observed pattern — sometimes called go-go, slow-go, no-go — has retirees spending more in the earlier, more active years (travel, hobbies, helping family) and less as activity naturally declines with age, with a partial uptick late in life from healthcare costs. A single static number averages over that whole shape and gets none of it exactly right — potentially overfunding the quiet later years while underplanning for an active start, or the reverse.
You don't need a perfect model of your own aging to benefit from this. Even a rough three-phase view — a higher early number, a lower middle number, a healthcare-aware late number — captures more of reality than one flat line ever will.
The number most people plug into a calculator isn't their real spending. It's their memory of their spending, flattened into a single figure that ignores how retirement spending actually moves over time.
The inverse question: what can your plan actually sustain?
Once real spending is on the table, the standard framing — "will $1.5 million survive?" — turns out to be the less useful question for most people, especially anyone still years from retiring. It bakes in an assumption (that $1.5M figure) that came from somewhere else, and it only tells you pass or fail on that one guess.
The more useful direction runs backward from what you already have: given your actual savings, what you're still contributing, your planned retirement age, and your expected Social Security, what's the highest monthly spending your plan can sustain at a confidence level you choose? That's an inverse calculation — instead of testing one spending number against your savings, it searches for the spending number your savings actually support.
This reframes the whole exercise usefully. Instead of "do I have enough for the number I picked," you get "here's what my actual plan supports" — a number that adjusts automatically as your real inputs change, rather than one you have to keep re-guessing.
Why "the number" isn't really one number
Even the inverse calculation doesn't produce a single fixed figure, because it depends on choices you get to make, not just facts about your savings:
- Retirement age. Retiring earlier means more years the portfolio has to cover and fewer years of contributions — the sustainable spending number moves with it, not always by a small amount.
- Social Security timing. Claiming earlier locks in a smaller check for life; waiting increases it. Since Social Security typically covers a meaningful share of retirement spending, shifting the claim age shifts how much the portfolio needs to carry on its own.
- Spending flexibility. A plan with room to cut back in a bad market can sustainably support a higher number in good conditions than a plan with zero flexibility, because it has a lever to pull if things go sideways.
- Confidence target. This is the one people underweight most. Asking "what can I spend and still succeed 90% of the time" produces a meaningfully different number than asking for 95% or 99% — pushing the confidence target up pushes the sustainable spending number down, because the plan is being asked to survive a wider range of bad outcomes, not just the typical one.
Four inputs, each one legitimately adjustable, each one moving the answer. That's why "the number" as a single fixed figure was never really the right way to ask the question.
A worked example
Take someone 50 years old with $1,000,000 saved, contributing $25,000 a year, planning to retire at 65, expecting $2,500/month from Social Security at full retirement age, and a 7% nominal expected return. Instead of asking "does my spending guess survive," the inverse calculation asks: what's the highest monthly spending this plan can sustain, at a chosen confidence level?
The answer isn't one number — it moves with the confidence level you're willing to accept:
| Confidence target | Read | What's different |
|---|---|---|
| 80% | More room, less margin | Highest sustainable spending, but a wider range of simulated futures fall short |
| 90% | A reasonable working target | Meaningfully lower spending than 80%, in exchange for holding up in more scenarios |
| 95%+ | Little room for surprises | Lower still — the plan is being asked to survive nearly every simulated bad stretch, not just the typical one |
None of these three numbers is "wrong." They're the same plan answering three different questions about how much safety margin to buy, and the tradeoff is visible instead of hidden — which is exactly what a single flat "retirement number" can't show you.
What this doesn't capture
An inverse calculation is still a simulation, not a promise, and it inherits the same honest limits as any projection. It can't know your actual future healthcare costs, a tax law that hasn't been written yet, or the specific sequence of returns markets will actually deliver — it prices in a wide range of historically plausible outcomes, not a forecast of the one that will happen. It's a projection, not a prediction. Treat the sustainable-spending number as a well-reasoned planning input, not a figure to spend to the exact dollar without checking back in as life changes.
How this is calculated
The number above comes from a 1,000-scenario Monte Carlo simulation that searches across possible monthly spending levels and finds the highest one that still meets your chosen confidence target, using randomized market returns based on historical data, your savings, contributions, retirement age, and Social Security timing. It's the same engine behind the full app — free to run, no account required. The complete math, including the specific assumptions and data sources, is documented in the methodology.
Common questions
This article runs one calculator on default numbers. The full app models your complete plan — taxes, Social Security optimization, spending shape over time, and stress-testing against real market history.
Run your full plan free →