How Compound Interest Actually Works — With the Arithmetic Shown

Compound interest gets described as the eighth wonder of the world roughly once a week, usually by someone selling something. It isn't a wonder. It's multiplication applied repeatedly, and once you've seen the arithmetic on paper it stops being mystical and starts being useful — including in the direction nobody puts on a motivational poster, which is what it does to a credit card balance.

Everything below is general education rather than investment advice, and every growth rate in it is an assumption chosen to make the arithmetic legible. Nobody knows future returns, and past performance does not predict future results. What the math does tell you is how the machine works, which is enough to make better decisions with.

The whole formula, once

Money that earns a return, and keeps the return, earns a return on the return. That's it. Written out: ending balance equals starting balance multiplied by (1 + rate) raised to the number of periods.

Take $1,000 at an assumed 7% a year, with nothing added and nothing withdrawn:

• After 10 years: $1,000 × 1.07^10 = $1,967
• After 20 years: $1,000 × 1.07^20 = $3,870
• After 30 years: $1,000 × 1.07^30 = $7,612
• After 40 years: $1,000 × 1.07^40 = $14,974

Look at what the decades produce. The first ten years add $967. The fourth ten years add $7,362 — more than seven times as much, on the same money, at the same rate, with no additional effort. The account didn't get smarter. It got bigger, and 7% of a bigger number is a bigger number.

That's the entire insight, and it explains a rule of thumb worth memorizing. Divide 72 by the annual percentage rate and you get, roughly, the number of years it takes for money to double. At 6%, about 12 years. At 8%, about 9. At 12%, about 6. The Rule of 72 is an approximation, but it's close enough to do in your head at a red light, and it works just as well on debt as on savings.

Why ten early years can beat thirty late ones

This is the part that changes behavior, so here it is with the arithmetic showing. Two savers, same assumed 7% annual return, contributions made at the end of each year.

Dana contributes $6,000 a year from age 25 to 35 — ten deposits, $60,000 out of pocket — then never adds another dollar. At 35 her balance is $6,000 × [(1.07^10 − 1) ÷ 0.07] = $82,899. That sum then compounds untouched for 30 more years: $82,899 × 1.07^30 = about $631,000 at 65.

Marcus starts at 35 and contributes the same $6,000 a year for thirty straight years — $180,000 out of pocket. At 65 he has $6,000 × [(1.07^30 − 1) ÷ 0.07] = about $567,000.

Dana put in a third of what Marcus did and finished ahead by roughly $64,000. She didn't earn a better rate or pick better investments. Her money simply spent more time in the machine. Under an assumed constant rate, time enters the formula as an exponent and the contribution enters as a multiplier, and exponents win.

Two honest notes on that example before anyone builds a life plan around it. Real returns are not constant — they arrive in a jagged sequence, and a bad stretch early or late changes the ending number. And $6,000 a year at 25 is not a realistic ask for a lot of 25-year-olds. The point is not that Marcus failed. It's that if you have any room at all right now, the years themselves are doing work that money can't buy back later.

Compounding frequency barely matters

Banks advertise daily compounding like it's a feature. Run the numbers on $10,000 at 5% for one year:

• Compounded annually: $10,500.00
• Compounded monthly: $10,511.63
• Compounded daily: $10,512.71

The gap between annual and daily is $12.71 on $10,000. Compare that to the gap between a 0.40% savings account and a 4.00% one — $360 on the same balance in the same year — and you can see where your attention belongs. Rate and time are the variables. Compounding frequency is a rounding error dressed up as a selling point.

The same machine, running in reverse

Nothing about this math cares which direction the money flows. A credit card compounds against you with exactly the same enthusiasm.

The Federal Reserve's G.19 consumer credit release reported an average rate of 22.15% on credit card accounts assessed interest as of June 2026, with about $1.3 trillion in revolving credit outstanding nationally. Take a $6,000 balance at 22.15%.

The monthly rate is 22.15% ÷ 12 = 1.846%. Month one, interest alone is $6,000 × 0.01846 = $110.75.

• Pay $150 a month and $110.75 of that first payment is interest. $39.25 touches the balance. At that pace the card takes about 74 months — a bit over six years — and costs roughly $5,000 in interest on a $6,000 purchase.
• Pay $300 a month and the same balance clears in about 26 months for roughly $1,550 in interest.

Doubling the payment cut the interest by about 70% and the time by two-thirds. That is not a discount anyone offered you. It is the exponent working the other way, and it's why a 22% balance sitting alongside a retirement account is a math problem before it's a psychology problem. Our guide to escaping high-interest debt and the walkthrough on paying off credit card debt both take that apart.

Where the poster version of this lies to you

Three things the motivational chart leaves out.

Inflation eats part of the gain. If a balance grows 7% in a year when prices rise 3%, the purchasing power gain is closer to 4%. Every long-horizon projection you have ever seen is in nominal dollars unless it says otherwise, and $631,000 in forty years does not buy what $631,000 buys today.

The assumed rate is doing enormous work and it is not a promise. Change Dana's 7% to 5% and her ending balance falls to about $326,000 — roughly half — from a two-point change in a number nobody can know in advance. Any calculator that hands you a confident ending balance is showing you arithmetic, not a forecast. The SEC's free compound interest calculator at investor.gov makes this explicit by letting you run a range of rates instead of one, which is the honest way to use it.

And the "start at 22 or lose everything" framing is the worst thing this math has been used for. If you're 48 and reading this with $9,000 saved, the correct response is not despair. You still have compounding years left, plus higher contribution limits than a 25-year-old and the catch-up provisions on top. For 2026 the IRS set the 401(k) elective deferral limit at $24,500 with an $8,000 catch-up from age 50 — and $11,250 instead for those turning 60 through 63 — and the IRA limit at $7,500 with a $1,100 catch-up. Our guide for people behind on retirement savings is built for exactly this.

What to do with this in the next fifteen minutes

1. Find your highest interest rate — usually a credit card — and divide 72 by it. That's how fast the balance doubles if you stop paying. Write the number down somewhere you'll see it.
2. Set one automatic transfer into a retirement or brokerage account, even at $25 a month. Automating it is what converts "time in the machine" from a nice idea into a fact.
3. If your employer matches 401(k) contributions, capture the full match before anything else in this article. That's an immediate return on the contribution that no compounding assumption competes with.
4. Run your own numbers at investor.gov's compound interest calculator using two rates — an optimistic one and a pessimistic one — and plan against the pessimistic one.

Then leave it alone. The formula's only real input that you control completely is how many years you let it run, and checking the balance weekly does not add any.