Standard deviation measures how much individual returns deviate from the average return over a period. For a stock with monthly returns, calculate the average monthly return, then measure each month's deviation from that average, square each deviation, average the squared deviations (variance), and take the square root to get standard deviation. Annualized standard deviation = monthly standard deviation × √12 (or daily × √252 for trading days). A stock with 20% annualized standard deviation has returns that fall within ±20% of the mean roughly 68% of the time (one standard deviation range), within ±40% roughly 95% of the time.

Historical annualized standard deviations as context: US T-bills ~1%, US aggregate bonds ~4-5%, US large-cap equities ~15-17%, small-cap equities ~20-23%, emerging market equities ~25-28%, individual large-cap stocks ~25-45%. These figures reflect long-term averages; realized volatility changes dramatically across market regimes. The VIX (CBOE Volatility Index) measures the implied 30-day volatility of the S&P 500 based on option prices — it spikes during crises (reaching 80+ in March 2020) and compresses during calm bull markets (dropping below 12-13 in 2017-2019).

Standard deviation's central limitation is symmetry — it treats upside and downside volatility identically. An asset that swings +30% followed by -10% has the same standard deviation contribution from both moves, even though the downside is what investors actually fear. For assets with positively skewed return distributions (where large gains occur occasionally but small losses are frequent), standard deviation overestimates the 'bad' risk investors experience. For assets with negatively skewed distributions (where occasional large losses occur — typical of options-selling strategies), standard deviation dramatically underestimates tail risk.

Standard deviation also assumes returns are normally distributed — the familiar bell curve. In practice, financial returns exhibit 'fat tails': extreme events (crashes and booms) occur far more often than a normal distribution predicts. The 2008 financial crisis, the March 2020 COVID crash, and the 1987 Black Monday each represented moves that were theoretically many standard deviations beyond the 99.99% probability range using normal distribution assumptions — but occurred in reality. This 'black swan' phenomenon means that risk models calibrated to historical standard deviation significantly underestimate actual tail risk in extreme scenarios.

Risk parity is a portfolio construction approach that weights assets to equalize their volatility contribution rather than their capital contribution. A traditional 60/40 portfolio has 90%+ of its volatility from the 60% equity allocation because equities are 3-4× more volatile than bonds — the bond allocation provides capital ballast but minimal risk ballast. A risk parity approach allocates capital proportionally to the inverse of each asset's volatility, then often applies leverage to bring total portfolio volatility to a desired target level.

Volatility targeting — scaling equity exposure inversely with realized volatility — systematically reduces risk in high-volatility regimes and increases exposure in low-volatility regimes. Research shows volatility-targeting strategies improve Sharpe ratios of underlying strategies by smoothing the risk contribution across time. The mechanism: during crashes, realized volatility spikes, the strategy reduces equity exposure automatically, limiting drawdown. During calm markets, low volatility increases equity exposure, capturing trend returns. This mechanical risk management is a form of systematic momentum applied to the volatility dimension.