How much of my paycheck should I be investing?

Ask the internet what share of your paycheck to invest and three confident numbers come back within the first page: 10%, because it is round; 15%, credited vaguely to experts; 20%, descended from a 2005 budgeting book’s rule of thumb. Each arrives without a footnote. None of them asks a single question about you first, which is a strange property for financial guidance, because the person earning $61,000 in a city where rent takes half of it and the person earning the same salary in a paid-off house are being handed the same number. A percentage that has never met your rent is not advice. It is a slogan with a percent sign.

This article takes the question seriously by running it in the opposite direction. Instead of starting from a target and commanding your month to comply, it starts from what months actually produce, in the government’s own data, and works toward the number your circumstances can defend. The stakes justify the ten minutes: the gap between a rate chosen honestly and a rate copied from a listicle is not cosmetic. Held for a career, a few points of pay compound into the difference between retiring comfortable and retiring anxious, and I will show you that arithmetic with real figures before the end.

What months actually produce

Start with the measured baseline, because it is nothing like the slogans. The Bureau of Economic Analysis tracks the personal saving rate, saving as a share of income after taxes, and the Federal Reserve’s FRED database carries the series back to 1959. As of May 2026, the national rate is 3.0%. Not 15. Three.

The gold band on that chart is the range the popular advice occupies, 10% to 15%, and the two tick marks are what Americans have actually done. Three things follow, each doing real work. First, the advice band sits far above the national behavior, which does not make the advice wrong, but does mean most people repeating it are describing a destination, not a norm, and you should stop grading yourself against an imaginary average. Second, the full record behind that chart runs wider than the band: FRED’s series bottoms at 1.4% in July 2005, spikes to 31.8% in the locked-down April of 2020, and averages 12.2% across the 120 months of the 1970s. The rate is a creature of circumstances: interest rates, housing costs, stimulus checks, fear. Normal moves. Third, and most useful: if the whole country’s rate swings with conditions, then yours legitimately can too, and a percentage is something you revisit as conditions change, not a vow you break.

The circumstances are not evenly distributed either, which is the quiet flaw in every one-number answer. The Bureau of Labor Statistics’ Consumer Expenditure survey for 2024 prices the average American household’s life at $78,535 a year against $104,207 of income before taxes, with housing alone taking $26,266, a third of all spending. And the spread around those averages is enormous: households in the lowest income fifth spent $35,046 while the highest fifth spent $150,342. The same survey finds households averaging $1,991 a year in contributions to retirement plans, which against average income is under 2% of pay. That is the honest landscape the confident percentages are shouted into: fixed costs that vary by multiples, and actual investing behavior far below every slogan. One definitional footnote for fairness: the saving rate counts every unspent dollar, parked and invested alike, so it makes the investing picture look better than it is rather than worse, and the BLS retirement-contribution figure is the sharper lens on the specific question this article is asking.

The number is an output, not an input

Here is the method this article defends, and it inverts the listicles. You do not adopt a percentage and then torture your month until it fits. You read your month, find what it can release, and the percentage is simply what that number happens to be when divided by your paycheck. The percent is the output of the arithmetic. The internet hands it to you as the input, which is exactly backwards, and is why so many people bounce between resolving to invest 15% and investing nothing.

Reading your month takes one evening and no app. Take home pay, minus the bills that arrive regardless, minus a realistic figure for the life you actually live, not the ascetic one you imagine under resolution conditions. What remains, after a starter cushion exists to absorb surprises, is the money your month can genuinely release every payday. For one person that computes to 4% of pay, and for another 22, and both numbers are correct, because correct means sustainable through your real months, and the full order those dollars should follow, match, expensive debt, cushion, then investing, has its own complete answer on this site.

While we are being precise, notice a question every slogan skips: percent of what? Fifteen percent of gross pay and fifteen percent of take-home are different numbers by a quarter or more, and the advice never says which it means. The conventions in the wild disagree too: workplace retirement plans deduct a percentage of gross salary, while a brokerage automation takes a dollar figure from the account your take-home lands in. This article’s method sidesteps the ambiguity, since you are finding dollars first and the percent is only a description. But you asked which one, so here is a plain answer rather than a shrug: use gross pay. It is the base workplace retirement plans already deduct against, it is the base the popular percentages were built on, and it is the harder of the two tests, so a rate that clears it cannot be flattered by the choice. Whichever you pick, keep it; a number that silently switches denominators will make you look better in exactly the way that costs money later.

One clause in that order refuses to be relative, though, and it outranks everything else here. If your employer matches retirement contributions, the percentage that captures the full match is your floor, whatever your month looks like, because matched dollars pay 50 to 100 cents on arrival and no other destination for a dollar comes close. Below the match, the right percentage is not a matter of opinion. It is arithmetic with only one answer.

What a single percentage point of pay is actually worth

None of which means the number is unimportant, and this is where the internet’s arguing about 10 versus 15 misses its own point. The difference worth studying is not between the slogans. It is between any number and that number plus a little, held for a long time, because on a career’s timescale a single percentage point of your pay, one point, is enormous.

Take a $61,000 salary, held flat for simplicity, and the difference between investing 5% of it and 10%: $254 against $508 a month. At a flat 8% a year, compounded monthly:

And note what the chart’s one unrealistic assumption does to its conclusion. The salary is held flat for thirty years, which no real career manages; the moment raises exist, a percentage quietly gives itself a raise too, since the same 10% of a growing number is a growing number. The flat chart is therefore the pessimistic version. Every promotion widens that gap further without a single new decision, which is the strongest argument for thinking in percentages at all once your dollars are found.

Five percentage points of pay, $254 a month, and the ending gap is $378,800, more than either path’s total contributions. That is the honest reason to care about your rate, and notice what it implies about strategy. The person who spends this year at 5% because that is what the month can hold, and reaches 10 over the next few years as raises land and debts die, captures nearly all of that gap. The person who declared 15% in January, met it twice, and quit in March captures none of it. On a thirty-year chart, the sustainable number you keep beats the impressive number you abandon, every time it is tried.

A word for everyone whose income refuses to arrive in tidy biweekly pieces, because percentage advice fails them worst of all. Freelance months, commission months, seasonal months: a fixed percent of an unfixed number produces chaos, and the dollars-first method is the escape. Set the automatic amount low enough that the schedule survives your quietest season without breaking, and treat whatever a strong month brings in above it as its own decision, made when the money is real rather than forecast. The schedule holds the floor; the good months volunteer the rest. That structure survives income weather that would shred any resolution phrased as a percentage.

So treat the percentage as a dial with a direction, not a grade. It should drift upward across your career as the forces that pinned it down release: incomes rise, debts end, cushions fill. The mechanics of moving it are almost embarrassingly small, a few minutes in a payroll portal or a brokerage automation, and each point moved is another six-figure line on that chart. When the dial should move, and what should trigger it, is a judgment your own months will announce; the trigger I trust most is any month that ends with money left over twice in a row.

The ceilings, for calibration

One more set of numbers belongs here, not as targets but as walls of the room. For 2026 the IRS raised the 401(k) contribution limit “to $24,500, up from $23,500 for 2025,” and the IRA limit “to $7,500 from $7,000.” Those are the official ceilings on tax-advantaged investing for an ordinary earner: $32,000 a year of sheltered room before any employer match, which on a $61,000 salary is more than half of gross pay.

Read them the right way around. The ceilings are not a suggestion that you should be filling them; against the BLS’s measured $1,991 average household retirement contribution, almost nobody is. They are the answer to a question that quietly stops many savers a few years in: is there a point where investing more stops being worth it, or where the good accounts run out? For practical purposes, no. The tax-advantaged room above nearly every real paycheck is vast, the room above it in an ordinary brokerage account is unlimited, and your rate can keep climbing for a whole career without hitting anything but your own months. The constraint was never the system’s capacity. It is, and will always be, what this article started with: the month, honestly read, revisited yearly, dialed upward when life allows.

Which is the entire answer, compressed: the match first, because its return is not optional arithmetic; then the dollars your real months can release, expressed as whatever percentage they happen to be; then a hand on the dial for the rest of your working life, moving it up as the months loosen, with $378,800 chapters waiting at every five points. The internet’s slogans are not evil. They are just answers to a question about somebody else’s month, and yours is the only one whose arithmetic you can actually run.