A withdrawal rate is an output, not something you choose
The familiar rule of thumb — one fixed percentage of the starting portfolio, adjusted upward each year for inflation — came from research on a specific historical market, a specific asset mix, and a fixed thirty-year horizon. It was never intended as a spending instruction. It was a test of whether a portfolio survived, not a description of how anyone actually lives.
The rule breaks in ordinary situations. It assumes your spending rises smoothly with inflation, when in reality it lurches: a roof, a car, a wedding, a parent's care. It assumes one horizon, when a couple has two. It assumes the money is in one pot, when most people hold a pre-tax account, a taxable account, maybe a Roth, and property. It says nothing about tax, which is the difference between a dollar of withdrawal and a dollar of spending.
So the number you want comes out at the end. First you work out what your spending is made of and where it can come from. The rate is the arithmetic that falls out of those answers, and it will not be the same number in year one as in year twelve.
Split your spending into the part that cannot move and the part that can
Every retirement budget has two layers. The floor is what you would still be paying if the market halved and you decided to change nothing else about your life: housing, insurance, food, utilities, medication, transport, the minimum you consider non-negotiable. Above that sits the flexible layer: travel, gifts, the second car, the renovation, meals out, help for adult children.
The useful question is not "what can I spend?" but "how much of the floor is covered by income that does not depend on markets?" Social Security, a pension, an annuity, rental income net of costs — that is your covered floor. Whatever the floor exceeds, your portfolio must cover, and that portion is the fragile part of the plan. A retirement where the portfolio funds only discretionary spending can absorb a terrible decade. A retirement where the portfolio funds the mortgage and the insurance cannot.
This is also where claiming decisions stop being a tax question and become a spending question. Claiming Social Security earlier reduces the monthly amount permanently, and delaying past full retirement age increases it — confirm the current reduction and credit figures with the Social Security Administration, since they are set by statute and formula rather than by your plan. That trade is really about how large a market-independent floor you want at eighty-five, paid for by drawing harder on the portfolio in your sixties.
The second half of the split matters just as much. Write down, now, what you would actually cut and by how much. Not the pious answer — the real one. If the honest answer is "nothing," you have no flexible layer, and your safe spending number is lower than any rule of thumb will tell you.
The average return will not happen to you — the order will
Two retirees with identical average returns over twenty-five years can end up in completely different places. The one who met a bad market in the first five years sold shares at low prices to pay for groceries, permanently shrinking the base that later growth compounds on. The one who met the same bad market at seventy-five had already banked a decade of growth. Same average, different life. This is sequence risk, and it is the single reason a plan that works on a spreadsheet fails in practice.
The practical consequence is that the first five to ten years of drawing money deserve their own design. That usually means holding some spending money in a form you are not forced to sell at the wrong time, and having decided in advance what happens if the portfolio drops sharply. The decision you want to avoid is the one made in the middle of the fall, at the worst possible moment, with no rule to point at.
Guardrails are the plainest version of that rule. You set a spending band. If the portfolio falls below the lower boundary, the flexible layer gets trimmed by a set amount. If it climbs past the upper boundary, some of the flexible layer is released. The point is not the specific band. The point is that the decision is made once, in a calm room, and it is written down. Almost nobody does this, and it is why so many people either underspend for a decade out of fear or keep spending straight through a downturn and never recover the base.
Where the money sits determines what a dollar of spending costs you
Two households with the same portfolio value can have very different safe spending, because tax is not a footnote — it is a claim on the same pot. A dollar out of a pre-tax retirement account is taxed as income. A dollar from a taxable brokerage account may trigger capital gains on only the gain portion, and the rate depends on your holding period and your income that year; the IRS sets out how gains are treated, and the rules are worth checking against the current guidance rather than memory. A dollar from a Roth account, if the conditions are met, is different again.
This creates a window that most people discover after it closes. Between the end of employment income and the start of required minimum distributions, taxable income is often at its lowest point for life. That gap is when partial Roth conversions or deliberate realization of gains can be cheapest. Required minimum distributions eventually force money out of pre-tax accounts whether you need it or not — confirm the current starting age and calculation with the IRS, because both have been changed by legislation more than once. If you spend the low-income years drawing only from taxable accounts because it feels tidiest, you may arrive at the forced-withdrawal years with a large pre-tax balance and a higher bracket than you ever had while working.
There is a second cost layered on top. Retirement-year income affects health insurance costs, and the mechanism differs depending on your age. Before Medicare, marketplace premium support is income-tested, so the size of a Roth conversion can change your premium in the same year. After Medicare begins, premium surcharges are also driven by reported income from earlier years. These are the joints where good advice on one decision quietly makes another decision worse — the investment person optimizes the portfolio, the accountant optimizes this year's return, and nobody owns the interaction.
Real spending is not a flat line, and it is not one person's line
Observed retirement spending tends to be highest in the early active years, sags in the middle as travel and activity fall away, and can rise steeply at the end if care is needed. A plan built on a flat inflation-adjusted line describes almost nobody. It usually underfunds the first decade — the years you were actually retiring for — and understates the tail risk of long-term care, which is the largest single uncertainty in most retirement budgets and the one least likely to be insurable on acceptable terms by the time people look.
The shape has hard edges early on too. Anyone leaving work before Medicare eligibility has to bridge health coverage, and that bridge is a real line in the budget, not a rounding error. Marketplace coverage is one route, and its cost depends on the income you report — which is also the income your withdrawal strategy produces. Coverage cost and withdrawal strategy are the same decision viewed from two sides.
Then there is the second person. For a couple, the plan has two horizons and one of them ends first. When one spouse dies, household spending usually falls — but rarely by half. Housing, insurance, and property costs are mostly unchanged. Meanwhile household income can drop sharply: one Social Security payment stops, and a pension may pay a reduced survivor amount or nothing at all. Survivor benefit rules depend on the claiming choices made years earlier, which is why a claiming decision made purely to maximize combined lifetime income can leave the surviving spouse with a thinner floor for what may be a long stretch of years. If you test your spending number against only joint life, you have tested the easier case.
The number changes, so build the review before you need it
A safe spending figure is a decision made under uncertainty, not a fact discovered once. What makes it hold is not precision at the outset — it is a scheduled point each year where you look at the portfolio, the actual spending, and the tax picture together, and adjust the flexible layer before drift becomes damage.
What separates the plans that hold from the ones that quietly fail is coordination. The gaps sit between decisions: between the claiming choice and the conversion window, between the conversion and the health insurance premium, between the withdrawal order and the survivor's floor, between the spending you set and the market you actually got. Each of those pairs is owned by nobody in a typical setup, because the investment adviser owns one, the tax preparer owns another, and the insurance decision was made in a different year by a different person.
The honest position: nobody can tell you your safe number without your actual balances by account type, your covered floor, your real floor-versus-flex split, your health situation, and your marital and survivor picture. Anyone who gives you a percentage before asking for those has given you a statistic, not an answer. What you can do is get the structure right, so that when a number is produced it means something — and so that you know in advance which lever you will pull when the market disagrees with your plan.
What to actually do
- Write out twelve months of actual spending from bank and card statements, then mark each line as floor or flexible. Do not estimate from memory; the gap between the estimate and the statements is usually large and always in the same direction.
- Total the income you will have that does not depend on markets — Social Security at your intended claiming age, pensions, annuities, net rental income — and subtract it from the floor. Whatever is left is the part the portfolio must cover no matter what, and it is the number that determines how much risk your plan can tolerate.
- List every account by tax treatment — pre-tax, Roth, taxable, HSA, property — rather than as one portfolio balance. Then map which years fall between the end of earned income and the start of forced withdrawals, because that window sets when conversions or gain realization are cheapest.
- Decide the rule now for a sharp market fall: which spending lines get trimmed, by how much, and at what portfolio level the trim triggers. Put it in writing and date it, so the decision exists before the day you need it.
- Test the same spending number against the survivor case — one Social Security payment stopping, the pension paying its survivor amount, and housing costs largely unchanged. If the plan only works while both of you are alive, you have found the weak point.
- Price the health coverage bridge if you are stopping work before Medicare, and check how the income your withdrawal plan produces affects that cost in the same year. Treat coverage and withdrawal sequencing as one decision, not two.
- Set a fixed annual date to review spending, portfolio, and tax position together, and name who is responsible for looking at all three at once. If the answer is nobody, that is the gap.
How this shows up
A couple retires at sixty-two with most of their savings in a 401(k). They claim Social Security immediately so they can leave the portfolio alone, which feels prudent. The claim permanently reduces both their monthly amount and the survivor amount, and it also raises their taxable income during precisely the years when converting some pre-tax money would have been cheapest. Eight years later, required withdrawals push them into a higher bracket than they ever occupied while working. No individual decision was careless. The sequence was never looked at as a whole.
A single retiree sets a spending figure using a fixed-percentage rule and a portfolio heavy in equities. Markets fall roughly a quarter in his second year. He keeps spending at the planned level because that is what the rule said, selling into the decline for eighteen months. Markets recover fully, but his base does not, because the shares that would have recovered were sold. Had he trimmed the flexible layer for two years — travel and a planned renovation — the outcome would have looked entirely different. The rule he followed had no instruction for a bad year.
A couple builds their plan around joint spending and never models the first death. When one spouse dies at seventy-eight, household income drops by one Social Security payment and the pension steps down to its survivor rate, while property taxes, insurance, and utilities stay where they were. The surviving spouse faces a fifteen-year stretch with a floor that no longer covers the floor.
Frequently Asked Questions
You can use one as a rough sanity check on whether you are in the right order of magnitude, and that is genuinely useful early on. It cannot serve as your spending plan, because it assumes flat inflation-adjusted spending, one horizon, one pot of money, and no tax — four assumptions that are false for nearly every household. The rate you can actually sustain depends on how much of your floor is covered by market-independent income and how much of the rest you would really cut in a bad year.
Observed spending patterns do tend to be higher in the active early years and lower in the middle, which is an argument for not forcing a flat line onto your plan. The cost is that early spending is also when sequence risk bites hardest, so front-loading raises the damage a poor first decade can do. The way people usually handle the tension is by keeping the early extra spending inside the flexible layer, so it can be cut rather than becoming a fixed commitment.
It changes who carries the risk rather than removing it. Converting part of your portfolio into contractual lifetime income raises your market-independent floor, which means the remaining portfolio supports only discretionary spending and can absorb a worse market — but you give up liquidity and control of that capital, the payments depend on the insurer's ability to pay and on state guaranty limits, and inflation protection is a feature you pay for separately if you want it at all. Whether that trade suits you depends on how large the uncovered part of your floor is.
They eventually force taxable money out of pre-tax accounts whether or not you need it that year, which can push you into a higher bracket and raise income-tested costs like Medicare premium surcharges. The starting age and the calculation have both been changed by legislation, so confirm the current rules with the IRS rather than relying on what was true a few years ago. The planning consequence is that the years before those withdrawals begin are usually the cheapest years to move money out of pre-tax accounts deliberately.
Getting it slightly wrong is normal and recoverable if you review it and adjust the flexible layer. What is hard to recover from is a wrong number combined with a poor first decade of returns and no willingness to cut, because you will have sold assets at low prices to fund spending you could have deferred. The asymmetry is worth noting: overspending early does lasting structural damage, while underspending early mostly costs you experiences you cannot get back later.
In most setups, nobody. The investment adviser manages the portfolio, the accountant files last year's return, and the insurance decision was made separately, so the interactions between them go unowned. If you want that coordination, it has to be someone's explicit job, and it is worth asking a prospective adviser directly how they handle withdrawal sequencing, conversion windows, and survivor income together.