Historical Volatility: Annualized Return Dispersion

Key Takeaways

• Historical volatility is the annualized standard deviation of log returns over a lookback window, scaled by the square root of 252 trading days.
• A 20-day daily standard deviation of 1.2 percent annualizes to roughly 19 percent, the number that makes cross-asset comparisons meaningful.
• Historical volatility describes the past and is not a forecast; volatility can change dramatically between a calm period and the week after a major shock.
• Confusing historical volatility with implied volatility is a common options error, they measure different things and frequently diverge significantly.

What It Is

Historical volatility (HV), also called realized volatility, is the standard deviation of an asset's returns over a lookback window, scaled up to a one-year figure. A stock with an HV of 25 percent has moved, on average, with a standard deviation of about 25 percent per year around its mean return during that window.

HV describes the past. It does not predict future movement, though it is often the starting point for doing so. It is computed directly from price data and requires no special market access, just a series of closing prices.

The Intuition

"Risky" is a slippery word until you put a number on it. Two stocks may both drift upward at similar average rates, but one may do so in a straight line while the other zigzags by several percent a day. Standard deviation of returns captures that wobble in a single number.

The annualization step matters because raw daily or weekly standard deviations are tiny and hard to compare. A daily standard deviation of 1.5 percent sounds small; its annualized equivalent of about 24 percent lines up with the long-run volatility of a typical large-cap stock, which is more intuitive. Annualizing puts every asset on a common scale regardless of the sampling frequency.

How It Works

The calculation has three steps.

Step 1. Compute periodic log returns from the closing price series:

r_t = ln(Close_t / Close_{t-1})

Log returns are preferred over simple percentage returns because they sum neatly across time and are symmetric around zero, which matters for the annualization math.

Step 2. Take the sample standard deviation of the return series over an N-period window:

sigma_period = stdev(r_t) over last N periods

N is a design choice. A 20-period window (roughly one trading month for daily bars) is a common default. Longer windows produce smoother estimates; shorter windows react faster to regime changes.

Step 3. Annualize by multiplying by the square root of the number of periods in a year. For daily data, the convention is 252 trading days:

HV_annualized = sigma_period * sqrt(252)

For weekly data use sqrt(52); for monthly data use sqrt(12). The square root appears because, under the standard assumption that returns are independent across periods, variance scales linearly with time and standard deviation scales with the square root of time.

Worked Example

Assume you compute daily log returns for a stock over the past 20 trading days and find a sample standard deviation of 1.20 percent, or 0.012 in decimal form.

Annualize with sqrt(252), which is approximately 15.87:

HV = 0.012 * 15.87 = 0.1905

The 20-day historical volatility is about 19.05 percent annualized. If the same stock two months later prints a 20-day standard deviation of 2.00 percent, its HV jumps to 0.02 x 15.87 = 31.7 percent. That increase signals a meaningful regime change, not a minor wobble.

Comparing HV across assets also works. An index with a 20-day HV of 12 percent is calmer than a single stock at 30 percent, which in turn is calmer than a small-cap running at 60 percent. The numbers give you a ranking that raw price charts obscure.

Common Mistakes

1. Confusing historical volatility with implied volatility. HV is computed from past prices. Implied volatility (IV) is backed out of current option prices and reflects the market's forecast of future volatility. They often differ. Options traders watch the spread between them as its own signal. Treating them as interchangeable is a common beginner error.

2. Mixing log returns and simple returns. Pick one convention and stick with it. Computing the standard deviation of simple returns and then annualizing with sqrt(252) gives a slightly biased answer because simple returns do not sum linearly. The industry convention is log returns for volatility work.

3. Choosing a window size by feel. A 5-day HV moves wildly; a 252-day HV barely moves at all. Different windows answer different questions. Short windows detect regime shifts; long windows estimate a baseline. State which one you are using and why.

4. Using calendar days instead of trading days. The sqrt(252) factor assumes you are counting trading days (about 252 per year in the US). If your return series skips weekends already, do not also divide by 365. Match the annualization factor to the sampling frequency.

5. Extrapolating HV into the future. Historical volatility is an estimate of what happened, not a forecast. Volatility clusters: quiet periods tend to be followed by quiet periods and vice versa. But regimes do change, and a 20-day HV measured the week before a crash will badly understate what is about to happen.

Frequently Asked Questions

Q: What is historical volatility in simple terms? Historical volatility is a percentage that tells you how much an asset's price has been moving, on average, over a past period, expressed as an annualized figure so you can compare stocks, indices, and other assets on the same scale.

Q: How does historical volatility affect investment decisions? It calibrates position sizing and stop distances. A stock with 30 percent HV moves about twice as much as one at 15 percent HV. Position sizing that does not account for this difference gives a misleading sense of risk in a portfolio.

Q: What is a real-world example of historical volatility? A large-cap index like the S&P 500 typically runs at 12–18 percent HV during calm markets. During the March 2020 COVID panic, its 20-day HV spiked above 80 percent. That signal warned risk models that normal stop distances and position sizes needed to be cut dramatically.

Q: How can investors use historical volatility practically? Use 20-day HV as a baseline for setting stops: if HV is at 25 percent, the typical daily move is about 1.6 percent, so a stop closer than 1.5 percent from entry will be hit by normal noise. A simple rule: size positions inversely to HV so that dollar risk stays constant across different volatility environments.

Q: How is historical volatility different from the VIX? Historical volatility is computed from actual past price changes. The VIX is implied volatility, derived from current S&P 500 option prices, and reflects the market's forward-looking expectation for volatility over the next 30 days. When VIX is much higher than HV, the market expects turbulence ahead.

Sources

1. Macroption. "Historical Volatility Calculation." https://www.macroption.com/historical-volatility-calculation/
2. AnalystPrep (FRM). "Measuring Return, Volatility, and Correlation." https://analystprep.com/study-notes/frm/part-1/quantitative-analysis/volatility/
3. SimTrade. "Historical Volatility." https://www.simtrade.fr/blog_simtrade/historical-volatility/
4. The Motley Fool. "How to Calculate Annualized Volatility." https://www.fool.com/investing/how-to-calculate/annualized-volatility/

Disclaimer

This article is educational content only and is not financial advice. Nothing here is a recommendation to buy, sell, or hold any security. Consult a licensed advisor before making investment decisions.