How to plan for retirement: the four questions that decide it

Retirement is a spending-and-assets problem, not a birthday. The four questions that actually decide it — and where each one gets answered.

Most people begin with a birthday. They pick an age, usually 65 because that is when Medicare starts, and then spend years hoping the money works out.

That is backwards. Retirement is a spending-and-assets problem, not a date. It happens when what you have — savings, plus guaranteed income, minus what health care costs you before Medicare — can fund what you spend, for as long as you live. The age is what falls out of that. It is an output, not an input.

This page is the map. It does not answer any of the four questions below; it tells you what they are, why each one matters, and where the answer lives. If you know nothing yet, read this first and then follow the links. This is the front door to the Retirement hub.

The four questions

Everything that decides a retirement is downstream of four things. In this order, because each one needs the one before it.

Notice what is missing: "how much have I saved?" That is not a question, it is a fact. It only becomes meaningful once question 1 tells you what the money has to do.

Question 1: What will you actually spend?

Not your salary. Not a percentage of it. The number you live on in a year, in today's dollars.

Everything downstream is a multiple of this figure. If you spend $60,000 a year, the portfolio you need is some multiple of $60,000, the gap Social Security leaves is measured against $60,000, and the tax you pay is driven by how you fund $60,000. Get it wrong by 20% and every other number in your plan is wrong by 20%. Nothing else in retirement planning has that kind of leverage, and nothing else is as easy to check — the answer is sitting in twelve months of bank statements.

Two things make it harder than it looks. First, retirement spending is not flat: it tends to be higher in the early, healthy years, dip in the middle, and rise again at the end when care costs arrive. Second, your personal inflation rate is not CPI. Health care has historically run hotter than general prices, and it is a bigger share of a retiree's budget than a worker's — so a single blended inflation number quietly understates what you will need in year 20. Our spending planner exists to make that divergence visible, and inflation and retirement explains why.

Once you have the spending number, the next step is turning it into a portfolio target — and that is where the honest disagreement begins. Divide by 4% and you get 25× spending. But the credible research does not agree on the divisor: Morningstar's forward-looking figure is 3.9%, while Bengen's revised historical worst case is 4.7%. They are not two estimates of the same thing to be split down the middle — they answer different questions, with different portfolios. How much is enough does the arithmetic; the 4% rule in 2026 explains why the two numbers disagree.

Question 2: What guaranteed income do you already have?

Your portfolio does not have to fund your whole life. It has to fund the gap — what is left after income that arrives whether markets rise or fall.

For most households that means Social Security, and possibly a pension. Both are worth more than they look, because both are guaranteed: a dollar of lifetime, inflation-adjusted income is doing a job no portfolio dollar can do, and it removes that dollar from the reach of a bad market. But be realistic about the size. The average retired-worker benefit is about $2,071 a month, and even the maximum at full retirement age is $4,152 — and that maximum requires a full career of earnings at the taxable ceiling. Social Security was built as a floor, not a replacement.

When you claim changes the floor permanently. Claim at 62, the earliest possible age, and a benefit based on a full retirement age of 67 is reduced by 30% for life. Wait past full retirement age and it grows by roughly 8% a year until 70. That is the single largest lever most people have, and it is free — see when to take Social Security and run it through the break-even tool.

One thing almost every guide still gets wrong: if you spent part of your career in public-sector work with a non-covered pension, the old rules that cut your benefit — the Windfall Elimination Provision and the Government Pension Offset — were repealed by the Social Security Fairness Act, retroactive to January 2024. If you last checked your claiming maths before that, check it again; it changed. And if your income is a private pension, assume it is not inflation-adjusted, because private-sector COLAs are rare. Pension lump sum vs annuity covers that trade.

Question 3: How do you get to Medicare?

Medicare starts at 65. If you stop working before that, you buy your own health insurance until you get there, and this is the single most underpriced line in early-retirement plans.

It got harder this year. The enhanced premium tax credits expired on 31 December 2025, which brought back the subsidy cliff: household income one dollar above 400% of the federal poverty line forfeits the entire premium tax credit. Not a taper — zero. For a single person that line is $62,600; for a couple, $84,600. An unsubsidised benchmark silver plan for a 60-year-old runs around $15,914 a year, so crossing that line by a dollar is a five-figure event.

The cruel part is what counts as income. Not just wages — a Roth conversion, a traditional IRA withdrawal, a capital gain, and a pension all raise the MAGI the cliff is measured against. So in the years before 65, the way you fund your spending and the price of your health insurance are the same decision. Two people with identical portfolios can pay wildly different premiums depending on which account they drew from.

Below the cliff there is still help — an income-based cap on the benchmark premium survived, it just got narrower and harsher. And below 100% of the poverty line the subsidy disappears again, so it is possible to crush your income too far. Read health care before Medicare and the ACA subsidy cliff is back, then see where your income lands with the ACA bridge tool.

Question 4: In what order do you withdraw, and what does it cost?

You have three kinds of money — taxable, tax-deferred, and Roth — and they are taxed completely differently on the way out. Which one you spend first, in which year, decides how much of your own money you keep.

The textbook order is taxable, then tax-deferred, then Roth. It is a reasonable default and it is frequently wrong, because it ignores the calendar. Between the day you stop working and the day Social Security and required distributions start, you are the only person deciding what your taxable income is. That window is where the tax planning happens: filling a low bracket deliberately, realising gains at the 0% long-term rate, or moving money to Roth — see the Roth conversion ladder. Skip the window and it closes for good.

It closes because of RMDs. Required minimum distributions begin at 73 for people born 1951–1959, and 75 for anyone born in 1960 or later — most people reading this. Almost every article quotes a flat 73. If you were born after 1959, you have two extra years of runway that nobody told you about. But at the end of it, the IRS starts choosing your income for you: a large pre-tax balance produces a forced withdrawal that lands on top of Social Security, drags more of the benefit into tax, and can lift your Medicare premiums permanently. The RMD planner shows what yours will look like.

There is one more timing trap worth knowing before you plan anything: Medicare's income-related surcharge, IRMAA, looks back 2 years. The tax return you file at 63 sets your premium at 65. So the pre-Medicare years are not free of Medicare consequences — they are exactly when Medicare is watching. A plain withdrawal plan puts the whole order together.

The thing that decides whether it works: flexibility

The four questions size the plan. One thing decides whether it survives contact with reality, and it is not the accuracy of your return assumption.

It is sequence risk — the order returns arrive in. Two retirees with identical average returns over thirty years can end in completely different places if one of them meets the bad decade first, because selling into a downturn turns a paper loss into a permanent one. It is the reason a plan can be arithmetically correct and still fail. Our sequence-risk stress test shows the same returns in two orders; sequence of returns risk explains it.

The defence is not a better forecast. It is the ability to spend less for a year or two after a bad one, to hold a cash buffer so you are not forced to sell, and — for some people — to earn a modest amount early on (part-time work). A retiree who can flex is worth more than a retiree with a better spreadsheet. That is the whole finding, and it is more useful than any withdrawal rate.

Where to start this week

1. Total last year's spending. That is question 1, and it takes an evening.
2. Get your benefit estimate from ssa.gov. That is question 2.
3. If you might stop before 65, price the bridge with the ACA bridge tool. That is question 3.
4. Then run the Retirement Checkup or Am I on track? to see the gap — and which lever closes it.

Key takeaways

• Retirement is a spending-and-assets problem. The date is an output, not an input.
• Spending is the number everything else is computed from. Start there, not with your balance.
• Guaranteed income shrinks the job the portfolio has to do. Social Security is a floor — the average benefit is about $2,071 a month.
• Retiring before 65 means buying health cover yourself, and the ACA cliff is back for 2026.
• RMDs start at 73 or 75 depending on your birth year — not a flat 73.
• Flexibility beats precision. Being able to adjust matters more than picking the exactly right withdrawal rate.

Educational only — not financial or tax advice.