Third-Party Research & Methodology Only

This section shares summaries of third-party academic research and descriptions of quantitative models. The content represents the findings of the original researchers, not the opinions or recommendations of Foxholm Financial. Foxholm Financial does not publish hypothetical or backtested performance metrics on its quantitative research pages. All content is restricted to methodology, signal construction, factor logic, and risk architecture. SEC rules require that investment advisers not present misleading performance data, and our methodology-only approach reflects that standard and the firm's fiduciary obligations.

Volatility

Risk Metric Statistical Concept

Volatility measures how much and how quickly an asset's price moves. It is the most fundamental concept in investment risk, used in nearly every area of portfolio management, options pricing, and financial regulation.

In everyday language, a "volatile" investment is one whose price swings widely. A savings account has near-zero volatility. A technology stock might have high volatility, with its price rising 3% one day and falling 4% the next. Volatility does not distinguish between upward and downward movements. It captures the magnitude of price changes in both directions.

Definition

In quantitative finance, volatility is most commonly defined as the standard deviation of returns (a statistical measure of how spread out values are from their average). A higher standard deviation means returns are more dispersed, indicating greater uncertainty about future outcomes.

Core Concept

Volatility = Standard Deviation of Returns

If a stock has an annualized volatility of 20%, its annual returns typically fall within 20 percentage points of the average return about two-thirds of the time (one standard deviation). About 95% of the time, returns fall within 40 percentage points of the average (two standard deviations).

Volatility is usually expressed as an annualized percentage, even when calculated from daily or monthly data. To annualize daily volatility, multiply by the square root of the number of trading days in a year (roughly √252 ≈ 15.87). To annualize monthly volatility, multiply by √12 ≈ 3.46.

Types of Volatility

There are two fundamentally different types of volatility, each answering a different question.

Historical (Realized) Volatility

Historical volatility measures how much an asset's price actually moved over a past period. It is calculated by collecting a series of past returns (daily, weekly, or monthly), computing their standard deviation, and annualizing the result. A stock with a historical volatility of 25% over the past year experienced annual price fluctuations of roughly that magnitude.

Historical volatility is backward-looking. It tells you what happened, but it does not guarantee the same level of price movement going forward. A stock that was calm for a year can become volatile overnight if new information arrives.

Implied Volatility

Implied volatility is the market's expectation of future volatility, extracted from the prices of options contracts. When investors expect larger future price swings, they bid up option prices, which increases implied volatility. The VIX index is the most well-known measure of implied volatility for the U.S. stock market.

Implied volatility is forward-looking but reflects expectations, not certainty. On average, implied volatility tends to be slightly higher than subsequent realized volatility. This difference is called the "volatility risk premium" and exists because investors are willing to pay extra for protection against uncertainty.

How Volatility Is Measured

The most common measurement is standard deviation, but several related statistics are used in different contexts.

Measure	What It Captures	Common Use
Standard deviation	Average distance of returns from their mean	The default volatility measure in most contexts
Variance	Standard deviation squared	Used in portfolio math and optimization models
Downside deviation	Volatility of only negative returns	Used in the Sortino ratio; focuses on harmful volatility
Average true range (ATR)	Average daily price range including gaps	Common in technical analysis and position sizing
GARCH (Generalized Autoregressive Conditional Heteroskedasticity) models	Time-varying volatility that clusters	Captures the tendency for volatile periods to follow volatile periods

The choice of measurement window matters significantly. A 20-day standard deviation captures recent volatility and reacts quickly to market changes. A 252-day (one-year) standard deviation smooths out short-term fluctuations and provides a longer-term perspective. Neither is "correct" in isolation; the appropriate window depends on the investment horizon and the question being asked.

Practical Example

Consider three investments over the same one-year period, all with identical average annual returns of 8%.

Investment	Annual Return	Annualized Volatility	Typical Monthly Range
Short-term bond fund	8%	3%	Roughly −1% to +2%
Balanced stock/bond fund	8%	10%	Roughly −3% to +5%
Single technology stock	8%	35%	Roughly −10% to +12%

All three earned the same return, but the experience of holding each was very different. The bond fund delivered steady, predictable results. The technology stock produced the same average outcome but with stomach-churning swings along the way. An investor who panicked during a 10% monthly drop might have sold at the worst time, turning temporary volatility into a permanent loss. This is why volatility matters even when long-term average returns are identical.

Volatility in Portfolios

One of the most important insights in modern portfolio theory is that a portfolio's volatility is almost always lower than the weighted average of its individual holdings' volatilities. This is the mathematical foundation of diversification.

The reduction happens because different assets do not move in perfect lockstep. When one asset falls, another may hold steady or rise, partially offsetting the loss. The degree of offset depends on the correlation (the statistical relationship) between the assets. Lower correlation means more diversification benefit and greater volatility reduction.

Diversification Effect

Suppose Asset A has 20% volatility and Asset B has 20% volatility. If they are perfectly correlated (correlation = 1.0), a 50/50 portfolio also has 20% volatility. If their correlation is 0.5, the portfolio volatility drops to about 17.3%. If their correlation is 0.0, it drops further to about 14.1%. The lower the correlation, the greater the volatility reduction.

This principle is why a well-constructed portfolio of many assets can deliver a given level of expected return at lower volatility than any single asset alone. It is the core rationale behind asset allocation and is formalized in the covariance matrix, which captures the relationships between all assets in a portfolio simultaneously.

Known Limitations

Limitations to Keep in Mind

Volatility is not the same as loss. A volatile investment can deliver strong long-term returns. An investment with zero volatility (like a savings account) may still lose purchasing power to inflation. Volatility measures uncertainty, not guaranteed loss.
Symmetric by design. Standard deviation treats a 5% gain and a 5% loss as equally "volatile." Most investors care more about downside movements than upside ones. The Sortino ratio addresses this by using only downside deviation, and the maximum drawdown focuses exclusively on worst-case losses.
Not constant over time. Volatility tends to cluster: calm periods are followed by calm periods, and turbulent periods are followed by turbulent periods. A stock with 15% annualized volatility over the past year might experience 30% volatility in the next month during a market shock. Robert Engle's ARCH (Autoregressive Conditional Heteroskedasticity) model (1982) and subsequent GARCH models were developed specifically to capture this clustering behavior.
Sensitive to the measurement window. A 30-day volatility reading can differ substantially from a 252-day reading for the same asset. Short windows react quickly to recent events but are noisy. Long windows are more stable but can miss emerging risk.
Normal distribution assumption is imperfect. Standard deviation assumes returns follow a bell curve. In reality, financial returns have "fat tails," meaning extreme events (both positive and negative) happen more often than a bell curve predicts. Benoit Mandelbrot's research in the 1960s first documented this property, showing that markets produce large moves far more frequently than standard models suggest.