Why the second half is a different problem
While you are saving, a market drop is at worst neutral and at best helpful. The balance falls, but you are still buying, and the units you buy during the drop are the ones that carry the recovery. Nothing is being sold at the bottom, so the loss is on paper until it is not.
Once you are living off the money, that reverses. Every withdrawal in a down year sells assets at a reduced price, and those units are permanently gone before the recovery arrives. The account has to make up not only the market decline but the extra shares that decline forced you to liquidate. This is why the accumulation plan cannot simply be run backward, and why the two phases deserve separate thinking even when they hold the same investments.
Sequence of returns, with the numbers
Take a $1,000,000 balance and three years of returns: minus 20 percent, minus 10 percent, and plus 40 percent. Two retirees get exactly those three numbers, in opposite orders, and each withdraws $50,000 at the start of every year.
| Year | Bad years first | Good year first |
|---|---|---|
| Start | $1,000,000 | $1,000,000 |
| Withdraw $50,000, then apply return | $950,000 at −20% → $760,000 | $950,000 at +40% → $1,330,000 |
| Withdraw $50,000, then apply return | $710,000 at −10% → $639,000 | $1,280,000 at −10% → $1,152,000 |
| Withdraw $50,000, then apply return | $589,000 at +40% → $824,600 | $1,102,000 at −20% → $881,600 |
| After three years | $824,600 | $881,600 |
A $57,000 gap after three years, from identical returns and identical withdrawals. Extend the pattern across a twenty-five or thirty-year retirement and the divergence compounds into a difference between comfort and running short.
Here is the part that makes the mechanism click. Run the same two sequences with no withdrawals at all: 0.8 × 0.9 × 1.4 gives 1.008 whichever order you multiply in, so both accounts end at $1,008,000 exactly. Order is irrelevant to a balance nobody touches. It becomes decisive the moment money is moving in or out, because then the return is being applied to a different amount of money each year depending on what came before. That single sentence is the whole of sequence risk.
You cannot control the order. What you can control is how exposed you are to a bad opening. The usual levers are holding enough in stable assets that the first stretch of withdrawals does not have to come out of whatever fell hardest; keeping some flexibility in the spending itself, so a bad year can be met with a smaller withdrawal rather than the planned one; and working, even part-time, in the early years. None of those is a product recommendation, and how to implement any of them for your situation is a conversation with a fee-only fiduciary advisor.
Withdrawal rates, and why the familiar one is contested
You will encounter a rule of thumb expressed as a fixed percentage of the starting balance withdrawn in year one and then adjusted for inflation each year afterward. It is worth understanding what it actually is: the output of a historical simulation over one country's market history, over a particular retirement length, with a particular portfolio mix, defining success as not hitting zero.
Every one of those is an assumption, and each has been argued over seriously.
| Assumption inside the rule | Why it is disputed |
|---|---|
| A fixed retirement length | Someone retiring at sixty in good health may be planning for thirty-five years, not thirty. The rate that survives one horizon may not survive the other. |
| One country's historical returns | The historical record it draws on is a single sample from a favourable period. Studies using broader international data generally produce lower sustainable figures. |
| A specific portfolio mix | Change the allocation and the answer changes, in both directions |
| Fees and taxes not fully accounted for | Investment costs and the tax on withdrawals come out of the same money, and neither is uniform across households |
| Spending that rises exactly with inflation | Real retirement spending is not smooth. It is usually higher early, lower in the middle, and can spike sharply late for care. |
| Success defined as not reaching zero | Ending with one dollar counts as success in the model and as a catastrophe in life |
The reasonable position is not that such rules are useless but that they are a starting frame with error bars, not a settled number. Anyone quoting a specific safe rate as established fact is skipping a live argument. Use the retirement income calculator to see how sensitive your own figure is to the assumptions, and re-run it every year against what actually happened rather than setting it once.
Thirty years of inflation, arithmetically
Inflation is the slow half of this problem and the one that is easiest to underplan, because nothing visible happens in any given year. Compounded across a retirement that may run three decades, it is not slow at all.
| Steady annual rate | Multiplier after 30 years | What $4,000 a month of spending becomes | Years to halve purchasing power |
|---|---|---|---|
| 2% | ×1.81 | about $7,250 | about 35 years |
| 3% | ×2.43 | about $9,710 | about 24 years |
| 4% | ×3.24 | about $12,970 | about 18 years |
At 3 percent, a household spending $4,000 a month at the start needs roughly $9,700 a month thirty years in to live the same way. A source of income that pays a flat, unindexed amount for life has lost well over half its usefulness by then, without ever missing a payment. The inflation value calculator will run the same conversion between any two years you name.
The practical response is to sort your income by whether it moves with prices. Social Security is adjusted periodically by a statutory mechanism, so it broadly tracks; the SSA publishes how and when. A private pension paying a flat amount usually does not, and it is worth asking that question specifically because the difference over thirty years is enormous. Portfolio withdrawals can be increased with inflation, but only if the portfolio supports it, which is exactly where sequence risk bites. Continued work tracks but is not available indefinitely. Anything paying a fixed nominal amount does not move at all, and the groceries do.
The exercise worth doing is writing your projected income into two columns and seeing how much of your monthly total sits in the column that does not move. Whatever is in that column is the part that quietly shrinks.
Health and care, the unknown that dominates
Every other uncertainty in a retirement plan has a bounded range. This one does not. Ordinary medical spending is largely a budgeting problem and can be estimated from coverage terms; what cannot be estimated is whether extended personal care will be needed, for how long, and at what level. The distribution is genuinely bimodal: many people need very little, and some need years of daily assistance at a cost that dwarfs everything else in the plan.
Quoting a national average here would be worse than useless, because almost nobody lands on it. What is worth stating plainly is that this is the largest single unknown most households face, that ordinary health coverage was not built to pay for long stretches of custodial care, and that the options for handling it — self-funding, insurance products of various kinds, family provision, and public programs with their own eligibility rules — all have significant tradeoffs and all are situation-specific. Find out what your actual coverage does, read Medicare's own materials rather than a summary of them for anything age-related, and put the question to a fee-only advisor and, where eligibility rules are involved, an elder-law attorney in your state. The practical, non-financial side of this is in the elder care starter guide, and the billing side is in the medical bill guide.
What you can still adjust after you stop working
The most powerful lever is also the most under-used: reducing the withdrawal in a bad year. Skipping one inflation increase after a poor stretch, or trimming discretionary spending temporarily, relieves the exact mechanism that causes sequence damage, and it costs nothing but flexibility. Part-time or seasonal work in the early years does the same job from the other side, shortening the drawdown at precisely the moment sequence risk is highest.
Housing is the largest single line for most households and the one where a change is possible but slow and emotionally expensive — see housing options. The timing of when benefits are claimed is consequential, permanent and rule-bound; take the version applied to your own record from the Social Security Administration and the tax interaction from a CPA. Which account you withdraw from first matters more than people expect because of tax treatment, and it is genuinely individual rather than a rule of thumb. And a cash reserve held outside the portfolio lets a bad market year be met without selling into it — the same logic as an emergency fund, applied to a household with no paycheck.
Reviewing it, rather than setting it
A drawdown plan set once and left alone is a forecast, and forecasts made thirty years out are not accurate. What works better is an annual review with three questions: what did I actually spend against what I planned, what did the portfolio actually do, and does the remaining balance still support the remaining years at the rate I am taking. Answering those honestly once a year catches drift while there is still room to respond. Answering them for the first time in year twelve does not.
Retirement income is often described as a solved problem with a known percentage. It is closer to steering than to arithmetic — a set of adjustments made in response to what the years actually deliver. The households that come through it well are rarely the ones that found the perfect rate; they are the ones that stayed willing to change the number.
Questions people ask
What is sequence-of-returns risk in plain terms?
It is the fact that when you are taking money out, the order in which good and bad years arrive changes the outcome, even if the average return is identical. Bad years early force you to sell assets cheaply, and those sold units are not there to participate in the later recovery. With no withdrawals the order makes no difference at all — multiplying the same returns in any sequence gives the same result. The risk exists only because money is moving out, which is exactly why the withdrawal phase needs its own planning rather than being treated as saving in reverse.
Is the 4 percent rule safe to rely on?
It is a starting frame with real assumptions inside it, not an established fact, and it is actively argued over. It came from simulation over one country's market history, for a particular retirement length, with a particular portfolio mix, and with success defined as not hitting zero. Change the horizon, the allocation, the fee and tax drag, or the data set used, and the sustainable figure changes — often downward. Use a rate as a planning input you test at several values, review it annually against what actually happened, and be prepared to adjust spending rather than treating any percentage as a guarantee.
How much should I assume for inflation over a long retirement?
Rather than committing to one number, look at what a range does to you. At 2 percent a year, prices roughly double in thirty-five years; at 3 percent, in about twenty-four; at 4 percent, in about eighteen. A household spending $4,000 a month at 3 percent inflation needs roughly $9,700 a month after thirty years for the same life. The useful exercise is sorting your income into what adjusts with prices and what does not, then looking at how large the fixed column is. That column is the part that quietly loses ground every year.
Which accounts should I withdraw from first?
This is genuinely individual and it is a tax question, not a rule of thumb. The right sequence depends on how each account is taxed, what other income you have in a given year, your filing situation, your state, and rules about required distributions that change and are set by statute. Getting it wrong costs real money over a long retirement, and getting it right can be worth more than a great deal of investment tinkering. Take this one to a CPA who can see your whole picture, and revisit it when your income situation changes.
How do I plan for long-term care costs?
Start by accepting that this is the biggest unbounded uncertainty in the plan and that most people underestimate it. Ordinary health coverage generally was not designed to pay for long stretches of custodial assistance, so find out precisely what your coverage includes rather than assuming. The available approaches — funding it from savings, insurance products, family arrangements, and public programs with their own eligibility rules — all involve significant tradeoffs and none is right for everyone. Read Medicare's own materials for anything age-related, and put eligibility questions to an elder-law attorney licensed in your state.