Prefer to watch? This post is also a 6-minute animated video, narrated by a raccoon with a candlestick tail:
Ask anyone what risk means and you’ll get something close to the textbook definition: the probability of a harmful event occurring, combined with the severity of its consequences. That’s the risk of security audits, insurance, aviation. By that definition a Ponzi scheme is high risk - decent probability of total loss, maximum severity. And the S&P 500 - five hundred of the most regulated, audited, litigated-against companies on earth, wrapped in one of the most supervised products in existence - reads as pretty safe.
Then those same people open a brokerage account and get told the S&P 500 is risk class 5 to 7 out of 7. Crypto, when it’s rated at all, pins the scale at 7. The word on the label is the word they use for fraud and engine failure, so they conclude the stock market is a casino with better marketing, and they leave their savings in cash. I’ve watched this exact dissonance stop smart people from investing at all.
Here’s the thing: finance never meant that risk.
What the label actually measures
The EU’s SRRI - the 1-to-7 number on every UCITS fund document, and the model for most “risk scales” you’ll meet - is defined by ESMA as a pure function of one input: the annualized standard deviation of returns over the last five years. The bands:
| Class | Annualized volatility |
|---|---|
| 1 | under 0.5% |
| 2 | 0.5% – 2% |
| 3 | 2% – 5% |
| 4 | 5% – 10% |
| 5 | 10% – 15% |
| 6 | 15% – 25% |
| 7 | over 25% |
No fraud probability. No counterparty quality. No regulation, no audit opinion, no severity term. A perfectly legal index fund and an outright scam could land in the same class, because the scale measures exactly one thing: how much the price wiggles. Risk, in finance, is volatility. The industry just kept the scarier word.
That mistranslation is why “high risk” reads as “gamble”. A gamble is negative expected value with a chance of ruin. A volatile asset is neither of those by definition - it’s just an asset whose price series has a wide standard deviation. Sometimes the two coincide. Usually they don’t.
Volatility is measured from inside the plane
There’s a second problem, and it’s sneakier: standard deviation of what, against what? Volatility needs a reference frame, and the frame is always fiat currency - money that is money because a government said so (fiat, Latin, “let it be done”), backed by nothing but the decree. We treat that frame as eternal, but it’s younger than the Boeing 747: the USD was redeemable in gold until 1971, when Nixon closed the gold window and a whole family of economic charts developed a kink - there’s an entire site, wtfhappenedin1971.com, dedicated to scrolling through them.
A reference frame has zero volatility against itself. That’s not a financial insight, it’s physics: a passenger in a smoothly flying plane with the blinds down measures their own speed as zero. Your cash sits in your account displaying the same number every day, so it feels like the fixed point of the universe - class 1, under 0.5%. But it’s class 1 for the same reason the passenger is “stationary”: you’re measuring the plane from inside the plane.
Step outside the plane and the picture changes. I pulled five years of ECB daily reference rates (2021-2026) and ran the numbers:
- JPY measured from an NZD reference frame: ~10-11% annualized volatility. That is SRRI class 5 of 7 - the same band as a global equity fund. (Fitting frame to pick, incidentally: New Zealand’s central bank invented inflation targeting in 1989.)
- JPY from a USD frame: ~10.4% annualized, class 5 again - and the dollar gained 47% against the yen over those five years. Two of the world’s major “safe” currencies, and holding one of them cost you nearly half its value in the other.
- EUR/USD: ~7.75%, class 4 - the band where a lot of bond funds live.
Nobody puts a “risk class 5” sticker on a Japanese savings account. But that’s only because the sticker is printed inside the plane. The yen didn’t get less volatile; you just weren’t allowed to see the window.
Volatility is also blind to direction
Standard deviation has no sign. An asset that grinds upward 30% a year with big swings and an asset bleeding 30% a year with the same swings get the same score. The scale cannot tell a rocket from a landslide - it only measures turbulence.
Bitcoin is the canonical example. I measured ~43% annualized volatility over the last year: class 7, off the top of the chart, maximum “risk”. Zoom out and the historical record says something the number can’t: people who held through a full cycle - four-plus years, through drawdowns that would make an equity investor physically ill - have overwhelmingly come out ahead. That’s not a promise about the future, and buying any top is real; but “class 7” and “you will probably lose money” turn out to be very different statements, and the label only ever made the first one.
The same asymmetry applies to boring old equity markets, and here the reference frame closes the loop. The central banks issuing the measuring stick target 2-3% inflation, on purpose, as policy. Whether you like that or not, it has an arithmetic consequence: even in a fantasy world where no company grew and nothing real changed, asset prices quoted in that stick should drift up a couple of percent a year, because the stick is designed to shrink. The market’s long-term upward bias isn’t only productivity and profit - part of it is just the ruler getting shorter. Volatility captures none of this. It penalizes the wiggle and ignores the slope, while the “riskless” alternative - cash - is the one asset guaranteed by explicit policy to lose purchasing power. On the actual definition of risk (probability × severity of harm), a savings account held for thirty years is not the safe option. It’s the certain, slow-motion loss.
Say volatility, mean volatility
I’m not arguing the 1-7 scale is useless. Volatility is a real, useful quantity - it tells you how rough the flight is, which matters enormously if you might be forced to get off mid-flight. If you need the money in two years, an asset that regularly drops 30% is genuinely dangerous to you, because your horizon can force you to sell at the bottom. That’s a real risk - but notice it lives in the interaction between the asset’s turbulence and your timeline, not in the asset alone.
What I’m arguing is that the word is doing damage. “Risk” imports probability-of-catastrophe intuitions into a number that measures wiggle amplitude. The fix costs nothing: call it volatility, because that’s what it is. Then reserve risk for the things that actually match the dictionary - counterparty failure, fraud, leverage, concentration, being forced to sell at the wrong time. A Ponzi scheme and an index fund differ enormously on every one of those. On a volatility band, they might not differ at all - which tells you everything about what the band does and doesn’t measure.
The next time a label says an index fund is “risk 6/7”, read it as what it is: turbulence 6/7, direction not included, as measured from inside a plane that’s been slowly descending by design since 1971. Less catchy. Considerably more honest.