Key takeaways
– Standard deviation (SD) quantifies how spread out numbers in a dataset are relative to the mean.
– In finance, SD is commonly used as a measure of historical volatility of returns.
– SD is the square root of variance and shares the same units as the original data.
– SD treats all deviation as “risk” (including upside), is sensitive to outliers, and relies on assumptions when combined with normal-distribution rules.
– You can compute SD by hand, with a spreadsheet, or with statistical software; for financial time series you commonly compute returns first and then annualize the SD.
Understanding standard deviation
Standard deviation measures the typical distance of individual data points from the dataset’s mean. If values cluster tightly around the mean the SD is small; if values are widely dispersed the SD is large. Mathematically
• Population variance (σ^2) = (1/n) Σ (xi − μ)^2
– Population standard deviation (σ) = sqrt(σ^2) = sqrt[(1/n) Σ (xi − μ)^2]
For sample data (estimating a population from a subset), use the sample variance with denominator (n − 1) and the sample SD: s = sqrt[(1/(n−1)) Σ (xi − x̄)^2]. Use the population form when you truly have the entire population of interest.
For price volatility (finance)
Applied to investment returns, SD indicates how much historical returns have varied around the average return—commonly interpreted as volatility. A high SD means returns have varied widely (higher risk); a low SD means returns have been steadier (lower volatility). Firms usually report SD for mutual funds and asset returns to communicate historical volatility.
For price trends
When comparing a fund to a benchmark, a small SD from the benchmark implies the fund closely tracks the index; a large SD implies greater deviation (active risk). SD also informs confidence intervals and hypothesis tests when returns are assumed to be roughly normally distributed.
Warning (what SD does and does not do)
– SD counts all variation as “risk,” including favorable deviations (above-mean returns).
– SD is sensitive to outliers, which can inflate the measure.
– Applying normal-distribution rules (68–95–99.7) requires the data to be approximately normal; many financial return series are skewed or fat-tailed.
– SD measures historical dispersion, not forward-looking probability of outcomes (unless you assume stationarity and other modeling assumptions).
Standard deviation formula (summary)
– Population SD: σ = sqrt[(1/n) Σ (xi − μ)^2]
– Sample SD: s = sqrt[(1/(n − 1)) Σ (xi − x̄)^2]
Step-by-step: Calculating standard deviation (manual)
1. List your data points: x1, x2, …, xn.
2. Compute the mean (population mean μ or sample mean x̄): mean = (Σ xi)/n.
3. Subtract the mean from each data point to find deviations: di = xi − mean.
4. Square each deviation: di^2.
5. Sum the squared deviations: Σ di^2.
6. Divide by n for a population variance, or by (n − 1) for a sample variance.
7. Take the square root of the variance to get SD.
Worked example (data: 5, 7, 3, 7)
1. Mean = (5 + 7 + 3 + 7)/4 = 22/4 = 5.5
2. Deviations: −0.5, +1.5, −2.5, +1.5
3. Squared deviations: 0.25, 2.25, 6.25, 2.25 → sum = 11.0
4. Population variance = 11/4 = 2.75 → population SD = sqrt(2.75) ≈ 1.6583
5. Sample variance = 11/(4 − 1) = 11/3 ≈ 3.6667 → sample SD = sqrt(3.6667) ≈ 1.9149
Key properties of standard deviation
– Always non-negative; zero only when all values are identical.
– Has same unit as the data (e.g., percent for returns).
– Not additive across datasets the way means are; variance of independent variables is additive but SD is not.
– Sensitive to scaling: multiplying all data by a constant multiplies SD by the absolute value of that constant.
Standard deviation vs. variance
– Variance = mean of squared deviations; has squared units (e.g., %^2).
– SD = square root of variance; returns the dispersion to the original units, making interpretation easier (e.g., percent volatility).
How standard deviation is used in business
– Risk management: quantify variability in portfolio returns, credit losses, or demand forecasts.
– Financial analysis: compare historical volatility across assets; evaluate risk-adjusted returns (e.g., Sharpe ratio uses SD in the denominator).
– Forecasting: assess uncertainty around forecasts (sales, cash flow) and build confidence intervals.
– Quality control: track process variation (e.g., Six Sigma uses SD-related metrics to measure defects).
– Project management: estimate variability in task durations (used in PERT and risk buffers).
Strengths and limitations of standard deviation
Strengths
– Uses all data points, giving a complete measure of dispersion.
– Widely understood and commonly reported; easy to compare across investments.
– Works with many statistical tools (confidence intervals, hypothesis tests).
Limitations
– Sensitive to outliers; a few extreme values can dominate SD.
– Treats upside and downside variation equally—does not distinguish “bad” from “good” volatility.
– Interpreting SD with normal-distribution rules requires the data to be approximately normal.
– Slightly more complex to compute manually than range or mean absolute deviation.
Tip (practical shortcuts)
– Quick estimate: if data are roughly normal, SD ≈ (max − min)/6 (because ±3σ covers most values). For a cruder estimate, some use range/4 for small samples.
– To annualize daily return volatility: σ_annual ≈ σ_daily × sqrt(252). For monthly: multiply by sqrt(12). (252 trading days is the common market convention.)
Examples of application: Apple share price volatility (how to compute)
Steps to compute historical volatility for Apple (AAPL):
1. Get historical closing prices for your period (daily, weekly, monthly).
2. Compute returns: simple returns = (Pt / Pt−1) − 1, or log returns = ln(Pt / Pt−1). Log returns are additive and commonly used in finance.
3. Compute the mean of returns.
4. Compute the SD of the returns (use sample SD if you treat the series as a sample).
5. Annualize the SD: σ_annual = σ_period × sqrt(number of periods per year). For daily returns use sqrt(252). Example: if daily SD = 1.5% → annualized ≈ 1.5% × sqrt(252) ≈ 1.5% × 15.874 ≈ 23.8%.
What does a high standard deviation mean?
– High SD indicates large historical dispersion around the mean—prices or returns frequently deviate widely from average. This suggests greater uncertainty and, in investing terms, higher volatility (and typically higher ex ante risk of large swings).
What does standard deviation tell you?
– SD quantifies typical variability around the mean. Combined with assumptions about the distribution (e.g., normality), it allows construction of confidence intervals and probabilistic statements (for example, roughly 68% of normally distributed returns lie within ±1 SD of the mean).
How do you find the standard deviation quickly?
– Use spreadsheet functions: Excel/Google Sheets:
• Sample SD: =STDEV.S(range)
• Population SD: =STDEV.P(range)
• Other functions: STDEVA, STDEVPA (include logical/text values differently).
– Or use a financial calculator or statistical software (R, Python’s numpy.std or pandas.Series.std with ddof parameter).
– Rule-of-thumb: SD ≈ range/6 for roughly normal data for a fast mental check.
Is lower standard deviation better in investing?
– Not necessarily. Lower SD means more predictable returns (less volatility), which is desirable for risk-averse investors. But some investors accept—and seek—higher volatility for the potential of higher returns. Portfolio decisions should weigh SD together with expected return and investor risk tolerance. Use risk-adjusted metrics (Sharpe ratio, Sortino ratio) to evaluate trade-offs.
Practical steps you can take right now
1. Decide whether your dataset is a population or a sample. Use population formulas for entire populations and sample formulas when inferring to a larger population.
2. In a spreadsheet, enter your data and use STDEV.S or STDEV.P as appropriate.
3. For financial returns, convert prices to returns first (preferably log returns), compute the SD of the returns, then annualize with sqrt(periods per year).
4. Check for outliers and skewness—consider trimming, winsorizing, or using other dispersion measures (e.g., mean absolute deviation or conditional volatility models) when appropriate.
5. When communicating risk to stakeholders, pair SD with interpretive context (expected return, confidence intervals, and a discussion of tail risk).
The bottom line
Standard deviation is a fundamental and widely used measure of dispersion that helps quantify volatility and uncertainty across many business and financial contexts. It is simple to compute with modern tools and valuable for comparing risk across assets or processes, but it has limitations—sensitivity to outliers, no distinction between upside and downside risk, and dependence on distributional assumptions—so use it alongside other measures and judgment.
Source
– Investopedia, “Standard Deviation,” Alex Dos Diaz .