Volatility: Standard Deviation, Annualization, and Investment Risk

What is Volatility?

Volatility describes how much investment returns vary over time.

In portfolio analysis, historical volatility is commonly measured as the standard deviation of periodic returns. A larger standard deviation means returns were more dispersed around their average during the measurement window.

A simplified sample standard-deviation formula is:

1s = sqrt[ Σ(r_t - r_bar)^2 / (n - 1) ]

where:

• r_t is the return in each period;
• r_bar is the average periodic return;
• n is the number of observed periods; and
• s is the sample standard deviation.

Volatility is therefore a dispersion measure. It is not a direct estimate of the maximum amount an investor can lose.

That distinction is the foundation for using volatility correctly.

A simple volatility example

Consider two hypothetical strategies with the same average monthly return.

1Strategy A monthly returns:
2+1%, +1%, +1%, +1%, +1%, +1%
3
4Strategy B monthly returns:
5+8%, -6%, +7%, -5%, +4%, -2%

Both series can produce similar average returns, but Strategy B moves much more widely around its average.

Its standard deviation is therefore much higher.

The difference matters because two investments with the same average return can create very different investor experiences and very different probabilities of short-horizon losses.

Volatility helps quantify that variation.

Volatility treats upside and downside variation symmetrically

Standard deviation squares each deviation from the average.

That means a large positive surprise contributes to measured volatility just as a similarly sized negative surprise does.

For example:

1Return 10 percentage points above the mean
2and
3Return 10 percentage points below the mean

contribute the same squared deviation.

This is useful when the analytical question is total return variability.

But it also creates an important limitation: many investors do not view unexpected gains and unexpected losses as equally harmful.

That is why Sortino Ratio uses downside deviation rather than total volatility, and why Maximum Drawdown focuses on the path from a portfolio peak to a later trough.

Different risk measures answer different questions.

Annualizing volatility

Daily, weekly, or monthly volatility is often converted to an annualized number.

Under the common assumption that periodic returns are independent and identically distributed, volatility scales with the square root of time.

For daily observations:

1Annualized volatility ≈ Daily standard deviation × sqrt(252)

For monthly observations:

1Annualized volatility ≈ Monthly standard deviation × sqrt(12)

Suppose the standard deviation of daily returns is 1.2%.

1Annualized volatility
2≈ 1.2% × sqrt(252)
3≈ 19.0%

That does not mean the investment is expected to gain or lose 19% in a year.

It means the observed daily return dispersion has been rescaled into an annualized standard-deviation convention.

The square-root-of-time rule becomes less reliable when returns are serially correlated, volatility changes through time, observations are smoothed, or the return process otherwise violates the simplifying assumptions.

Measurement frequency changes the answer

Volatility is not a single permanent property of an investment.

A reported number depends on choices such as:

• daily versus weekly or monthly returns;
• the start and end dates;
• the length of the lookback window;
• simple versus logarithmic returns;
• whether returns are gross or net of fees;
• whether stale or missing observations exist; and
• the annualization convention.

A strategy that appears calm using monthly marks can still experience large intramonth swings.

Illiquid investments are a particularly important example. Infrequent appraisals or stale prices can mechanically suppress measured return variability even when underlying economic risk is substantial.

Always ask how the volatility number was produced.

Historical volatility is backward-looking

A standard deviation calculated from past returns describes the observed sample.

It does not guarantee that future volatility will match the past.

Market regimes change.

Leverage changes.

Portfolio holdings change.

Correlations change.

A quiet historical window can be followed by a turbulent period, and a crisis-heavy sample can overstate what later becomes a calmer environment.

Therefore:

1Historical volatility is evidence about past dispersion,
2not a guaranteed forecast of future risk.

Forward-looking volatility estimates can be built from models or option prices, but those are different analytical objects with their own assumptions.

Volatility is not maximum drawdown

Suppose two strategies have similar standard deviation.

One may experience frequent small losses and gains.

The other may be stable most of the time but suffer one severe crash.

Their volatility can look similar even though their worst peak-to-trough experiences are very different.

Maximum Drawdown measures the largest observed decline from a previous portfolio peak to a subsequent trough.

Volatility instead summarizes dispersion across the return observations.

Neither measure dominates the other.

A disciplined performance review often considers both.

Volatility is not beta

Beta measures sensitivity to a chosen market or benchmark factor.

Volatility measures total return dispersion.

A security can have high idiosyncratic volatility but modest beta if much of its movement is unrelated to the benchmark.

Conversely, a diversified portfolio can have a high beta to the equity market even if its residual, security-specific risk is small.

The distinction can be summarized as:

1Volatility -> how much the investment varies
2Beta       -> how the investment co-moves with a benchmark factor

They are related in many portfolio models, but they are not interchangeable.

Volatility and the Sharpe ratio

The Sharpe Ratio uses volatility as its risk denominator.

A common form is:

1Sharpe Ratio = (Portfolio Return - Risk-Free Rate) / Volatility

That makes the Sharpe ratio a measure of excess return per unit of total measured return variability.

If two portfolios earn the same excess return, the one with lower volatility will have the higher Sharpe ratio, all else equal.

But the result inherits the limitations of the volatility estimate.

Changing the return frequency, lookback period, fee treatment, or annualization method can change the ratio.

This is one reason risk-adjusted ratios should be compared only when their underlying methodology is reasonably consistent.

Volatility can hide sequence risk

Standard deviation does not preserve the order of returns.

Rearranging the same set of periodic returns leaves the standard deviation unchanged.

Yet return order can matter greatly to an investor who is withdrawing capital or facing leverage constraints.

A deep loss early in retirement can be more damaging than the same loss after years of gains.

A leveraged strategy can be forced to deleverage during a drawdown even if its long-run standard deviation later looks ordinary.

Path-dependent measures such as drawdown help reveal risks that volatility alone does not capture.

Low volatility does not automatically mean low risk

An investment can display low measured volatility and still carry substantial economic risk.

Examples include:

• illiquid assets marked infrequently;
• short-option strategies that collect steady premiums before occasional large losses;
• concentrated credit exposures with rare defaults;
• leveraged positions protected by temporary market calm; and
• strategies whose historical sample excludes a relevant stress regime.

Likewise, high volatility is not automatically undesirable.

An investor with a long horizon may willingly accept substantial short-term price variation in exchange for a different expected-return profile.

Volatility is a measurement, not a quality score.

How investors should use volatility

A useful volatility review asks:

1. What return frequency was used?
2. What lookback period was used?
3. Is the number historical, forecast, or implied?
4. Was the volatility annualized, and under what convention?
5. Are prices liquid and frequently observed?
6. Is the return distribution strongly skewed or fat-tailed?
7. What do downside and drawdown measures show alongside standard deviation?
8. Is benchmark sensitivity better described by beta or tracking error?
9. Has the portfolio's composition or leverage changed materially during the sample?

For systematic-strategy context, Grizzly Bulls' Models research can be evaluated with these distinctions in mind. The Macroeconomic Conditions Index provides separate regime context, but it is not a live volatility calculation for this encyclopedia page.

Sources and further reading

• CFA Institute, 2026, Portfolio Mathematics: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/portfolio-mathematics
• CFA Institute, 2026, Portfolio Risk and Return: Part I: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/portfolio-risk-return-part-1
• CFA Institute, 2026, Active Equity Investing: Portfolio Construction: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/active-equity-investing-portfolio-construction
• CFA Institute, 2026, Measuring and Managing Market Risk: https://www.cfainstitute.org/insights/professional-learning/refresher-readings/2026/measuring-managing-market-risk

Explore more topics in the Financial Research Encyclopedia.