Saving for retirement stops being vague advice the moment you ask how much is enough. Nobody hands you that number. You build it out of what you spend, and it arrives as a range rather than a single figure. Here is how the range gets built and what actually moves it.
Start With What You Spend, Not What You Earn
The usual shortcut says to plan on replacing 70 or 80 percent of your income. It hides the two biggest reasons your costs fall the day you stop working.
You stop saving for retirement, because you are in it. You also stop paying payroll tax on earnings. The employee share of Social Security and Medicare tax is 7.65 percent of wages, and it comes out of paychecks, not out of retirement account withdrawals or Social Security benefits.
Work an illustrative case. You earn $80,000. You put $12,000 into a retirement account, pay about $6,100 in payroll tax and about $9,000 in income tax. You live on roughly $52,900. That is the figure your retirement has to reproduce, and it came out of arithmetic instead of a rule of thumb.
Other costs move both ways. A paid-off mortgage and no commute lower the bill. Health coverage before Medicare, and care costs later, raise it.
Where the 4 Percent Figure Actually Comes From
In October 1994, a financial planner named William Bengen published a paper in the Journal of Financial Planning called "Determining Withdrawal Rates Using Historical Data." He took 30-year retirements beginning across the decades of available US market history and ran each against the real record, using the S&P 500 for stocks, intermediate-term Treasury bonds for bonds, and the consumer price index for inflation.
His question was narrow. What is the largest first-year withdrawal, raised each year to match inflation, that never emptied the portfolio inside 30 years? In that data, about 4 percent. His worst case was someone who retired in 1968 and met falling markets followed by 1970s inflation.
Four years later, three professors at Trinity University in Texas published a similar test in the journal of the American Association of Individual Investors. Cooley, Hubbard and Walz varied the stock and bond mix and the length of retirement, then reported a success rate for each combination instead of one safe number. That paper is why the idea spread, under the name Trinity study.
Turning a Withdrawal Rate Into a Target
A withdrawal rate becomes a savings target with one division. Taking 4 percent means the balance is 25 times the withdrawal, since 1 divided by 0.04 is 25.
Use the illustrative $52,900 from above. At 4 percent, the target is about $1.32 million. Change the rate and the target moves hard. At 5 percent the multiplier is 20 and the target is about $1.06 million. At 3.5 percent it is about 28.6 and the target is about $1.51 million.
A figure taken from a study of one country's past is doing nearly half a million dollars of work in that comparison.
What Social Security Does to the Number
Your savings do not have to cover all of your spending. They cover the gap between your spending and the income that arrives whether or not markets cooperate.
Say Social Security pays $24,000 a year, again illustrative. Your savings now fund $28,900 instead of $52,900. At 25 times, the target falls from about $1.32 million to about $723,000. Nearly half of it vanished, and you did nothing but count income you had already earned.
That income has two features a portfolio does not. It rises with inflation each year, and it never runs out. Claiming later raises it permanently, which shrinks the savings target again. How the benefit is calculated, what claiming at 62 versus 70 does to it, and what the projected shortfall means are all covered in our article on Social Security. A pension or an annuity does the same arithmetic job.
Why the Order of Returns Matters as Much as the Average
While you are saving, a bad year early does little damage, because little money is in the market and decades of compounding lie ahead. While you are withdrawing, that flips.
Picture two illustrative retirees. Each starts with $1 million and takes $40,000 a year, and their returns average the same over ten years. One gets the bad years first, the other gets them last. The first sells into falling prices, so every withdrawal takes a bigger slice of what is left, and the good years that follow work on a shrunken balance. The second grows the balance before the withdrawals bite.
Same average, different order, and the two end up far apart. This is sequence of returns risk. Bengen's 1968 retiree is exactly that case, since the average return across those years was not the problem. The order was.
The Rest of the Case Against Calling It a Rule
Two further objections are serious, and neither is fringe.
It rests on one country's history. The researcher Wade Pfau applied Bengen's method to other developed countries' market histories and found that a 4 percent inflation-adjusted withdrawal was far less safe outside the United States. Same method, same 30-year horizon, different market data, and the answer changed. The twentieth century United States was an unusually good place to invest, and the finding partly measures that. A rule built on one country's unusually good century is a narrower fact than the name suggests.
It also assumes behavior nobody has. The method spends the same real amount in a crash year as in a boom year, for 30 years, without ever looking at the balance. Research by David Blanchett found that real spending tends to drift down through retirement and then rise late, mostly for health costs. A retiree who trims in a bad year is running a safer plan than the one the studies tested.
Your Number Is a Range You Revise
Every input above is a guess about a life you have not lived. Your spending at 65, how long retirement lasts, what Social Security pays, what markets do, what care costs. Precision here is false comfort.
Treat the output as a band instead. Run the arithmetic at 3.5 percent and at 5 percent and you have a low and a high. Revise it when something real changes: a raise, a move, a marriage, a child, a diagnosis. Later the choice stops being yours anyway, since pre-tax accounts eventually force withdrawals, which our article on required minimum distributions covers.
The Three Levers That Move It
Three things change the answer: spend less, work longer, save more.
Spending less is the strongest, because it works twice. Cutting $5,000 a year from your costs removes $5,000 from what your savings must produce, which at 25 times is $125,000 off the target. It also frees $5,000 a year to contribute.
Working longer is close behind, and most people underrate it. A 2018 paper written for the National Bureau of Economic Research compared the two directly. Delaying retirement by three to six months did about as much for a household's sustainable standard of living as saving an extra one percent of pay for 30 years. Extra working years add contributions, subtract withdrawal years, and let a larger Social Security benefit start later.
Saving more is the lever most advice leads with. Late in a career it is the weakest of the three, and that same paper found an extra percent saved over the final ten years matched working one extra month. Early it is the strongest, because compounding then has decades to work on it, which our article on compound growth walks through.
