What Is Volatility? Why Your Average Return Isn't What You Got
A fund gains 50%, then loses 50%. The average return is zero. You're down a quarter.
By Pavel Penev, MScFounder, TradeWize · 10+ years trading the marketsYou put $10,000 into a fund. In its first year it gains 50%. In its second year it loses 50%.
Add those two years up and divide by two. The average return is 0%.
Now count the money. The gain took you to $15,000. The loss took half of that away, so you're sitting on $7,500. You're down $2,500, which is 25% of what you started with.
Nobody charged you a fee. Nothing was mis-sold. The average really is 0%, and the $2,500 is really gone. That gap between the average and the money is what this article is about.
The short answer
Volatility is how much a set of returns bounces around its own average. It says nothing about direction — a fund that swings hard upwards is volatile too. What it does decide is how much of that average you keep. Every gain has to repair the loss before it, and repairing a loss takes a bigger gain than the loss itself. So the more the returns bounce, the further your money finishes below the average you were quoted.
What the volatility number actually measures
When a factsheet prints a volatility figure, it's printing a standard deviation. That sounds worse than it is.
Take a fund's yearly returns. Work out the average. Then work out how far a typical year sat from that average. That distance is the standard deviation, and it's the whole idea. A small number means the years huddled close to the average. A big number means they scattered.
Say a fund averages 10% a year with a volatility of 16%. One standard deviation either side of the average runs from a 6% loss to a 26% gain. So the fund telling you "10% a year" is really telling you its ordinary year lands somewhere in a 32-point range. Losing money is inside that range.
The same fund, drawn out. The average is the thin line in the middle. The shaded band around it is one standard deviation either side — a 6% loss at one end, a 26% gain at the other. Both halves belong to the same "10% a year".
You'll often read that about two years in three land inside that band. That number comes from the bell curve, not from any market. Real returns have fatter tails than a bell curve does, which means the extreme years turn up more often than the curve predicts. So treat two-in-three as a rough guide and check the actual record where you can.
Here it holds up. Of the 98 years of S&P 500 returns from 1928 to 2025, 65 landed inside their own one-standard-deviation band. That's 66.3%.
The band is wide, though, and that's the part that matters. Over those 98 years the average is 11.86% and the volatility is 19.40%. That puts one standard deviation either side at a 7.55% loss and a 31.26% gain. 26 of the 98 years lost money. And only 7 of the 98 finished within 2 points of that 11.86% average. The average year is a year the market almost never has.
Two funds, the same average, very different money
Here are two funds. Both average exactly 8% a year over 20 years. Put $10,000 into each and leave it alone.
Steady earns 8% every single year. Wild alternates: it gains 38%, then loses 22%, then gains 38% again, all the way through. Add Wild's 20 years up and divide by 20 and you get 8.00% — the same as Steady, to the decimal.
Watch one pair of Wild's years. Your $10,000 gains 38% and becomes $13,800. Then it loses 22%, and 22% of $13,800 is $3,036, so you're left with $10,764. Two years at an average of 8% should have added 16%. You got 7.64%.
| Steady | Wild | |
|---|---|---|
| What it does each year | Earns 8%, every year | Gains 38%, then loses 22%, over and over |
| Average annual return | 8.00% | 8.00% |
| Volatility | 0.00% | 30.78% |
| What actually compounded | 8.00% | 3.75% |
| $10,000 after 20 years | $46,610 | $20,880 |
Read the first two number rows and the funds are twins. Read the last two and one of them ends with 2.2 times the money. The only thing that changed in between is how much the returns bounced.
After 20 years Steady is worth $46,610 and Wild is worth $20,880. That's $25,730 less on an identical average. Nobody took the $25,730. It never existed, because Wild's average was never the rate your money grew at.
Both lines start at $10,000 and both funds average 8.00% a year. Steady climbs. Wild spends every second year giving back a chunk of the year before, and finishes $25,730 behind.
Why the gap exists: losses need bigger gains to undo
The reason sits in one piece of arithmetic, and it catches almost everybody.
A loss needs a bigger gain to reverse it than the loss itself. Fall 10% and you need 11.11% to get back to level, not 10%. That looks like a rounding quibble. It isn't, because it gets worse quickly.
| If you fall this far | You need this gain to break even |
|---|---|
| 10% | 11.11% |
| 20% | 25.00% |
| 30% | 42.86% |
| 50% | 100.00% |
| 75% | 300.00% |
| 90% | 900.00% |
Lose half your money and a 50% gain doesn't return you to level. You have to double.
The cause is that the gain works on a smaller pot. Lose 50% and the next gain is calculated on the half you still have, not on what you started with. The further down you go, the harder the money that's left has to work, and the curve steepens the whole way.
That's the drag. It's not a fee and nobody collects it. It's the shape of the arithmetic. Look at Wild again: undoing its 22% loss takes a 28.21% gain, so a big slice of every 38% year goes on repairs before a cent of it is growth.
This isn't a quirk of made-up funds. The worst year in the whole 98-year S&P 500 series is 1931, down 43.84%. Getting back to level from there took a 78.06% gain.
Learn it by doing
Reading about it is one thing — it clicks when you do it. Practise this hands-on in a free, interactive lesson (Stage 17: Portfolio-Level Risk).
Try the free lesson →The one-line formula for how much volatility costs
There's a shortcut for the size of the drag, and you can do it on a phone.
Take the average return. Subtract half the square of the volatility. What's left is roughly the rate that compounds.
average − (volatility × volatility) ÷ 2
Run it on Wild. Its volatility is 30.78%, which is 0.3078 as a decimal. Multiply that by itself and you get 0.0947. Halve it and you get 0.0474, or 4.74 percentage points. Take that off Wild's 8.00% average and the formula says it should compound at about 3.26%.
Wild actually compounded at 3.75%. The real drag was 4.25 points and the formula predicted 4.74, so the shortcut came out 0.49 points too harsh.
It's an approximation, and it drifts
The formula is close, not exact, and it gets looser the more volatile the thing is. At Wild's 30.78% it's 0.49 points out, which is a lot when the answer itself is 3.75%. Few real funds bounce anywhere near that hard. At ordinary levels of volatility it's much tighter, and the next section runs it on a real index where it misses by 0.05 of a percentage point. Use it to size the problem. Don't use it to plan a number.
Does it hold up on real market data?
Here's the S&P 500's annual total return with dividends reinvested, 1928 to 2025. That's 98 years.
Add those 98 years up and divide by 98 and the average is 11.86%. The rate that actually compounded over the same stretch is 10.02%. The gap between them is 1.84 percentage points.
Now the formula. The volatility of those 98 years is 19.40%. Half its square is 1.88 percentage points. The real gap is 1.84 percentage points. The shortcut is out by less than a twentieth of a percentage point across almost a century, and nothing here was fitted or tuned to make that happen.
The flat line is the 11.86% average, held still. The curve underneath is what the formula says you keep as volatility rises. The shaded space between them is the drag. The marked dot is the real index at its real 19.40% volatility. Nothing about the index went into drawing that curve.
Two things to be clear about
First, this uses US data because it's the longest clean run of annual figures anyone publishes, not because the effect is American. The drag is arithmetic. It applies to any market, any asset and any currency, and it always has. Second, 10.02% is what happened, not what's coming. Nothing here says the next 98 years will resemble the last 98. The century shows you the relationship between the two numbers. It predicts the size of neither.
What volatility doesn't mean
Volatility gets used as a synonym for risk. It isn't one.
Volatility is how much something moves. Risk is the chance you end up permanently worse off. They overlap a lot, and they come apart at both ends.
Say a fund swings hard, then recovers, and you never had to sell it. That cost you movement and nothing else. A company that falls 90% and never comes back costs you the money, and undoing that would take a 900% gain that's never arriving. Both of those show up as "volatility" on a factsheet. Only one of them is the thing you're actually afraid of.
It runs the other way too. Quiet isn't safe. A bond fund can post mild, low-volatility year after mild, low-volatility year and still hand you a real loss when interest rates move. What drives that loss is how long the fund's bonds run, not how much they wobbled. Our piece on duration walks through it. And a fraud is perfectly calm right up until the morning it isn't — the returns look wonderful and the volatility looks tiny, because somebody is typing the numbers in.
One more. A volatility reading measures the past, and a quiet stretch doesn't promise another one. Bollinger Bands are built straight out of standard deviation, and they make this visible. The bands pinch shut when a market has gone still, and that pinch often comes right before a big move. They don't say which way the move goes. Our piece on Bollinger Bands covers what they do and don't tell you.
Realized volatility, implied volatility, and the VIX
Everything so far measures movement that already happened. Take the returns, work out how far they sat from their average, done. That's called realized volatility, and it's a fact about the past.
There's a second kind, and it points the other way. When people buy and sell options, the price they settle on contains an assumption about how much the underlying thing will move before the option expires. Work backwards from the price and you can read that assumption out. That's implied volatility. It's the market's own guess at the movement ahead, and people are betting real money on it.
The VIX is implied volatility for the whole S&P 500 rolled into one number. The Cboe calculates it live from S&P 500 option prices, and it covers the 30 days ahead. It's quoted on the same annualised scale as everything else here. A VIX in the high teens means traders are bracing for about as much movement as the index itself has shown over the 98 years above.
It's a thermometer, not a compass. A high VIX says traders expect a big move and says nothing about which way — some of the largest up days on record landed in the middle of the worst panics.
If you buy options, implied volatility stops being background and becomes the price tag. It's why the same protective option costs several times more in a panic than it did the week before. It's also why an option can lose money on a correct call, once the event everybody was waiting for is over. Our piece on the options Greeks handles that properly.
What to actually do with this
None of this is an argument against volatility. It's the reason stocks have paid more than cash over long stretches. You're being compensated for putting up with the movement, and if the movement went away the compensation would go with it. The arithmetic changes three practical things.
Horizon. Volatility does its real damage when you're forced to sell in the middle of it, because that's the moment a temporary fall becomes a permanent one. Money you need in two years and money you need in twenty aren't the same money, and only one of them can afford a bad decade.
Contributions. Paying in monthly is a different experience from putting a lump in once, because a falling market buys you more shares per payment. That doesn't cancel the drag and nothing does. It does mean a bumpy decade treats a regular saver differently from someone who bought everything at the top. Our dollar-cost averaging piece gives the honest version, including where it doesn't help.
Diversification. This is the one worth the most. People call diversification the only free lunch in investing, and this article is the reason why. Mixing holdings that don't move together cuts a portfolio's volatility without cutting its average return. Cut the volatility while the average stays put and, by the arithmetic above, more of that average survives to compound. You're not being clever about what to pick. You're just paying less of the drag. Our piece on diversification covers how far that goes and where it stops.
The asterisk on every "7% a year" projection
If you've read our compound interest article, or any article like it, you've seen 7% a year, after inflation, used to grow a pot forward. That's a reasonable thing to do, and it's worth knowing whether a rate like that is an average or a compounded one. An average annual return can't be compounded forward honestly. Over these 98 years the average is 11.86% and the rate that compounded is 10.02%. Project with the first and you overstate the answer, and the bumpier the thing is, the more you overstate it.
Is volatility the same as risk?
No, though they're related closely enough that the words get swapped. Volatility measures how much returns move around their average, in both directions. Risk is the chance you end up permanently worse off. A fund that swings wildly and recovers costs you movement. A company that falls 90% and never returns costs you the money, and it would take a 900% gain to undo that. Volatility also misses the risks that are quiet by nature, like a bond fund's exposure to rising rates, or a fraud reporting invented returns.
What is a good volatility number?
There's no single good number. Volatility only means something next to two other things: what you're comparing it against, and what return came with it. A reference point helps. The S&P 500's own volatility over the 98 years from 1928 to 2025 is 19.40%. A fund quoting a lot more than that bounces harder than the whole US stock market has, and a fund quoting far less is usually holding bonds or cash alongside its stocks. Read the volatility and the return together. A low figure sitting next to a low return just means you're being paid less to hold it.
Why is my return lower than the average return?
Because an average and a compounded rate are two different calculations, and whenever returns vary at all the average is the higher one. The average adds the years up and divides. The compounded rate is what your balance actually grew at. Over the 98 years of S&P 500 data in this article the average is 11.86% and the compounded rate is 10.02%. That's a gap of 1.84 percentage points, caused by nothing but the movement in between. Fees and the timing of your own contributions sit on top of that.
How is volatility calculated?
Take the returns for each period. Work out their average. For each period, find the difference from that average and square it. Add those squared differences up, divide by one less than the number of periods, then take the square root. The result is the standard deviation, and that's the volatility. Squaring is what stops the good years and the bad years cancelling each other out. Every deviation counts, whichever side it fell on. Dividing by one less than the count is the convention factsheets use, and it's how the 19.40% above was worked out. Most of them annualise too, so a 16% volatility means 16% of movement a year.
Does high volatility mean the price is going to fall?
No. There's no direction in it at all. Volatility measures the size of the moves, not their sign, so a very volatile investment is one that could move a long way either way. High readings do tend to appear while markets are falling, because that's when everyone is trading nervously, but a violent rally produces high volatility too. Anything claiming to read direction off a volatility number is reading something that isn't in it.
Can you avoid volatility drag?
Not directly, because it isn't a fee anyone charges — it's what happens when returns vary at all. Holding something less volatile reduces the drag, and it usually reduces the return along with it, so that's a trade rather than a fix. The one thing that genuinely improves the trade is diversification. Mixing holdings that don't move together lowers a portfolio's volatility without lowering its average return, so more of the average is left to compound. Longer horizons help too, not by removing the drag but by giving you the option not to sell into it.
The average return is the easy half
Picking what to buy gets all the attention. How much a portfolio moves, what a bad year does to it, how much of the headline return survives the ride — that's the other half. It's the half that decides whether you're still holding in ten years, and it's the half you can actually measure.