How much do you need to retire — and how much can you safely spend?

Start from spending — your real number, not a guess

Every credible retirement estimate is built on annual spending, and almost everyone guesses theirs wrong — memory-based budgets routinely miss 20–30% of actual outflow (insurance paid annually, the property tax bill, the quiet subscriptions, the cars that get replaced every eight years). Before any multiplication, get 12+ months of actual spending from actual data across every account, then adjust it for retirement: subtract what disappears (payroll taxes, college, the mortgage if it ends), add what appears (healthcare before Medicare — routinely $15,000–$25,000/yr per couple in premiums and out-of-pocket — plus travel and the hobbies the job was suppressing).

Households with complexity have a second layer: which entity pays for what, income that arrives lumpy (RSU tails, K-1 distributions, rental income), and spending that isn't really spending (transfers between your own accounts). If the spending number is fiction, every downstream number is fiction with more decimal places.

The 25× starting point — and what it actually assumes

The standard shortcut: multiply net annual portfolio-funded spending by 25. It's the inverse of the "4% rule" — the finding (Bengen's work and the Trinity study) that an initial 4% withdrawal, adjusted for inflation each year, historically survived 30-year U.S. retirements at a stock-heavy allocation. Spend $200,000 a year beyond Social Security and rental income, and 25× says roughly $5M of portfolio does the job; at a more conservative 3.5% initial rate it's about $5.7M, and at 3% about $6.7M.

Know what the shortcut assumes: a ~30-year horizon (retire at 45 and you should study lower rates), U.S. historical returns (the future needn't cooperate), disciplined behavior in drawdowns, and — critically — that the portfolio number is measured after tax, which it never is on a dashboard. It's a screening tool, not a verdict: useful to know whether you're at 60% or 105% of a plausible target.

Taxes bend the number more than fees do

Two portfolios of $5M are not the same $5M. Traditional 401(k)/IRA dollars owe ordinary income tax on the way out — at a 30% blended rate, $2M of pre-tax money spends like $1.4M. Roth dollars spend at face value. Taxable dollars sit in between: only the gains are taxed, at capital-gains rates, and basis matters. A household with $5M concentrated in pre-tax accounts may truly hold $3.8M of spendable wealth; the same headline number weighted to Roth and high-basis taxable might spend like $4.7M. Any serious retirement estimate must be computed on after-tax spending power by account type — and the withdrawal order across those accounts (plus Roth conversions in low-income years, and required minimum distributions later) shifts lifetime taxes by six figures for large portfolios. The mechanics are knowable; the optimal sequence for you is CPA territory.

Once retired: safe spending is a system, not a number

The decumulation question — "how much can I safely spend?" — has a subtlety the accumulation one doesn't: sequence-of-returns risk. Two retirees with identical portfolios and identical average returns can end wildly differently if one meets a deep bear market in years one through five while withdrawing. Early losses plus fixed withdrawals is the one combination that breaks retirements, which is why the answer is a system rather than a single rate.

The systems worth knowing: fixed-real (the classic 4% mechanism — simple, but blind to markets), fixed-percentage of current balance (never depletes, but income swings with markets), and guardrails (start near 4–5%, cut spending modestly after bad years and raise it after good ones — the family of approaches, like Guyton-Klinger, that most planners actually use). A cash buffer of one to three years of spending is the common companion so a bad year's withdrawals don't come from depressed assets.

Stress-test it — don't trust a single simulation

Monte Carlo analysis runs your plan against thousands of return sequences and reports a survival rate. Used honestly, it's a stress test: 85% success doesn't mean "15% chance of ruin," it means "in 15% of simulated paths you'd need to adjust — spend less, work longer, or annuitize a floor." The assumptions (return, volatility, inflation, horizon) drive everything, so insist on seeing them, run pessimistic variants, and re-run yearly as reality replaces assumption. In Formation, the projection shows its drawdown math and assumptions explicitly, and every such figure is labeled an estimate — this is education, not a Formation forecast or recommendation.