Correlation-Adjusted Sizing
Correlated positions can turn several small bets into one big one. Learn how correlation adds hidden risk, simple adjustment rules and portfolio methods.
Correlation adjusted sizing means reducing position sizes when your trades are likely to move together. Five separate positions each risking 1% look like 5% total risk spread across five ideas. But if all five are technology stocks, or all long the US dollar, they can lose together on the same day, behaving like one 5% bet on a single idea. Adjusting for correlation keeps your real risk where you think it is.
How correlation hides risk#
Correlation measures how closely two assets move together, from +1 (perfectly together) through 0 (unrelated) to minus 1 (perfectly opposite). See Covariance and Correlation.
With low correlation, losses on some positions are offset by gains or smaller losses on others. With high correlation, losses stack.
Simple adjustment rules#
You do not need complex mathematics to manage correlation. Practical rules many traders use:
| Rule | Example |
|---|---|
| Group correlated positions and cap the group | No more than 2% total risk in any one sector or theme |
| Count highly correlated positions as one | Two positions with correlation above 0.7 share one position's risk budget |
| Reduce size for each additional correlated trade | Second correlated trade at half size, third at a quarter |
| Limit same direction exposure in one currency | Long EUR/USD, GBP/USD and AUD/USD together are largely one short dollar bet |
See Portfolio Heat and Concentration Risk.
A slightly more formal approach#
For two positions with risk amounts R₁ and R₂ and correlation ρ, combined risk (as standard deviation) is:
Combined risk = √(R₁² + R₂² + 2 × ρ × R₁ × R₂)
With ρ = 1, combined risk is R₁ + R₂ (no diversification). With ρ = 0, it is √(R₁² + R₂²), lower than the sum. With ρ = minus 1, the positions can offset each other.
For whole portfolios, this extends to a covariance matrix. See Portfolio Construction and Risk Contribution and Risk Decomposition.
Correlations change#
Correlations are not fixed. In calm markets, many assets look loosely related; in crises, correlations often jump towards 1 as everything is sold at once. A portfolio that looks diversified on normal days can behave like one large position in a crash. For that reason, many traders assume higher correlation than recent data shows when setting limits. See Correlation Management.
Common correlation traps#
- Sector clustering: many "different" stocks from the same industry.
- Index plus components: holding an index ETF and its largest members.
- Currency overlap: several pairs sharing the same base or quote currency.
- Crypto: most altcoins move closely with Bitcoin, especially in sharp sell offs.
- Hidden factors: stocks in different sectors that all depend on low interest rates or the same economic theme.
Common mistakes#
- Counting the number of positions rather than the number of independent bets.
- Using correlations from calm periods to size for stressful ones.
- Forgetting about direction: long one asset and short a correlated one can reduce risk, while long both increases it.
Frequently asked questions#
Why does correlation matter for position sizing?#
Because correlated positions tend to lose at the same time, so their combined risk is larger than each position's individual risk suggests.
How do I adjust for correlated positions?#
Group correlated trades and cap their combined risk, or treat highly correlated positions as a single position in your risk budget.
Do correlations stay the same?#
No. They change over time and often rise sharply during market stress.
Next, learn where to place the stops your sizing depends on: Stop Loss Strategies.
3 quick questions on this lesson. Get them all right to finish it.
Turn on JavaScript to take the quiz.
Mentioned in
- Fractional KellyRisk Management
- Position TradingStrategies and Styles
- Currency CorrelationsForex
- Portfolio ConstructionPortfolio and Performance
- Correlation and Beta CalculatorCalculators