Harbor Coffee, our fictional coffee business, has a reported beta of 0.7. Does being below one protect it from a sharp fall?
No. A stock can have a beta below one and still swing more widely than its benchmark.
Beta measures sensitivity to a market. Standard deviation measures the size of all the swings. A small first number does not cancel a large second one.
Volatility measures all the swings
Return volatility is how much returns vary from period to period. Standard deviation is a common measure of it: how widely those returns spread around their average. Use percentage returns, not dollar share prices. A higher share price does not make an investment more volatile.
Suppose investments A and B have the annual returns shown here. The calculations below use three annual USD total returns: income included, before fees, personal taxes and inflation adjustments.
Both average 0%. Yet A's sample standard deviation is 10 percentage points and B's is 20. B's bars reach twice as far from zero: same center, twice the variability. The average hides how rough the ride was.
A 10-point standard deviation describes the spread of returns, not a 10% loss or a 10% gain. And a zero arithmetic average does not mean you broke even; your compounded return answers that.
Beta measures the market connection
Beta estimates a holding's sensitivity to a specified benchmark from their historical returns. The benchmark's beta relative to itself is 1.
Use B as our market benchmark and add a stock, C, over the same three years:
| Annual return | Year 1 | Year 2 | Year 3 |
|---|---|---|---|
| Benchmark B | −20% | 0% | +20% |
| Stock C | 0% | −30% | +30% |
C's sample standard deviation is 30 percentage points, versus B's 20. Their correlation, which describes how their returns move together, is 0.5.
C's beta is 0.5 × 30 ÷ 20 = 0.75. Its standard deviation is 50% larger than the benchmark's, yet its beta is below one.
Year 2 makes the distinction concrete: C loses 30% while B goes nowhere. A low beta leaves plenty of room for trouble that belongs to the company alone.
In this fitted relationship, a −10% return for B corresponds to −7.5 percentage points for the part of C's return associated with B: 0.75 × −10%. Other influences can make C's total return very different.
These three-year examples teach arithmetic; they cannot support a forecast. A negative beta describes a tendency to move against the benchmark in the sample, not insurance against a falling market.
Compare like with like
For standard deviation, compare within the same asset class, using the same return frequency and lookback period, the span of history used. There is no universal safe percentage.
| Measure | Tells you | Reference | Misses |
|---|---|---|---|
| Standard deviation | Total variability | Matching samples | Future loss size |
| Beta | Market sensitivity | 1 for the benchmark | Company shocks |
Monthly and annualized volatility use different scales. Converting monthly volatility to an annual scale commonly uses a square-root-of-time shortcut. It assumes stable volatility and no correlation between different months' returns, and it ignores compounding.
For a whole portfolio, measure the portfolio's returns. Averaging its holdings' standard deviations leaves out how they move together.
A positive beta below 1 means weaker sensitivity than the benchmark's own; above 1 means stronger sensitivity. One is a comparison point, not a safety boundary.
All three inputs to the beta formula need matching dates and return frequency. Our formula gives a raw beta. Some providers adjust it toward 1, so the methodology matters too.
Can two "three-year betas" use different samples? Fidelity's fund comparison guide specifies monthly returns over 36 months. Our example uses three annual returns. Same span, but 36 observations versus three.
Harbor's 0.7 is missing its benchmark, lookback and frequency. You cannot assume it measures sensitivity to the S&P 500. Before using the number, you need to know what it was measured against and over which dates.
Neither number is a loss limit
Standard deviation counts upside and downside deviations from the average equally. Beta describes only the benchmark relationship. Neither tells you whether a business will fail, whether you can sell when you need cash, or how deep and long the next loss will be. Investment risk is bigger than either statistic.
A short history can miss bad conditions; a long one can describe a company that has since changed. A calm past does not set a limit on future losses. Returns can fall more than one standard deviation below their average.
To see how painful a past loss was, look at how far the investment fell below its peak and how long it stayed below it. That is what drawdowns examine.
In short
- Standard deviation measures the spread of all observed returns, including gains.
- Beta describes sensitivity to one specified benchmark.
- A stock can have a beta below one and still swing more widely than the market.
- Match the sample and units before comparing numbers. Neither statistic sets a maximum loss.