Volatility Targeting Formula: Position Sizing Without Guesswork
Volatility targeting converts a risk budget and a recent volatility estimate into a repeatable exposure rule. The formula is simple; the timing, leverage cap and trading policy are where most implementation errors occur.
Direct answer: volatility-targeted exposure is usually calculated as target volatility ÷ estimated volatility, subject to a leverage cap. If the target is 10% and the current estimate is 16%, the rule assigns 62.5% exposure. The arithmetic is easy; a credible implementation must also specify the estimator, lag, cap, rebalance rule and costs.
The formula
Both volatility inputs must use the same units and annualization convention. With daily returns, a common research convention estimates daily standard deviation over a fixed window and multiplies it by the square root of 252. That convention is an approximation, not a law of markets.
Volatility target calculator
This calculator returns a notional exposure multiplier. It does not model margin rules, borrowing costs, taxes, liquidity, gaps or investor suitability.
Worked example
| Target vol | Estimated vol | Raw multiplier | With 1.0× cap |
|---|---|---|---|
| 10% | 16% | 0.625× | 0.625× |
| 10% | 10% | 1.000× | 1.000× |
| 10% | 8% | 1.250× | 1.000× |
A cap changes the strategy. Without one, a very low estimate can create large leverage precisely when the estimator is most likely to be overconfident. The cap should therefore be part of the original rule, not added after seeing an uncomfortable backtest.
Five decisions that matter more than the division
1. Choose a volatility estimator before testing
A 20-day rolling standard deviation reacts faster than a 60-day window but may trade more often. An exponentially weighted estimate emphasizes recent observations. Compare a small set of defensible choices; do not search hundreds and report only the smoothest history.
2. Lag every input
Today's closing return cannot determine an exposure that supposedly earned today's return. A daily backtest normally applies an estimate available after one close to a position entered no earlier than the next executable decision point.
3. Declare leverage and exposure bounds
Long-only cash portfolios may cap exposure at 1.0×. Futures portfolios can exceed that level, but margin, financing, gap risk and instrument limits become part of the result. A floor can also be used to avoid shrinking positions into immaterial fragments.
4. Use a rebalance buffer
Rebalancing every tiny change can turn estimation noise into turnover. A threshold rule—trade only when the desired multiplier moves far enough from the current one—can reduce costs, but the threshold must be tested net of the same cost assumptions used elsewhere.
5. Separate risk control from return prediction
The formula attempts to stabilize risk exposure. It does not say whether the asset will rise. A lower multiplier can reduce both a future loss and a future gain, and a volatility spike can reverse before the estimate catches up.
What the evidence does and does not say
Moreira and Muir study portfolios that reduce factor exposure when recent realized variance is high. Their published results report improved risk-adjusted performance across several equity factors and currency carry in the examined samples. That is evidence about a defined historical strategy, not a guarantee for every estimator, asset or implementation.
Harvey and coauthors report an important qualification: the Sharpe-ratio improvement in their study is concentrated in risk assets such as equities and credit, while the effect is much smaller for bonds, currencies and commodities. They also find that volatility scaling reduces the severity of extreme outcomes more broadly. Together, the papers argue for testing the mechanism by asset class rather than treating volatility targeting as a universal return enhancer.
Common backtest mistakes
- Look-ahead exposure: using a same-day volatility estimate against the same-day return.
- Uncapped quiet-period leverage: allowing a low estimate to create unrealistic exposure.
- Free rebalancing: ignoring spread, market impact, financing and futures rolls.
- Risk-mismatched comparison: comparing a scaled strategy with an unscaled benchmark without equalizing risk.
- One favorable estimator: hiding sensitivity to the lookback, decay factor or rebalance threshold.
A reproducible specification
- Name the tradable instrument and return series.
- Define the volatility estimator, annualization and minimum observations.
- Lag the estimate to the first executable rebalance time.
- Set the target, cap, floor and rebalance buffer.
- Model turnover, financing and instrument-specific costs.
- Report gross and net results, drawdown depth and duration, exposure history and parameter sensitivity.
Frequently asked questions
What is the basic volatility targeting formula?
Divide target annualized volatility by estimated annualized volatility, then apply a pre-declared exposure cap and any minimum exposure rule.
Does volatility targeting guarantee a lower drawdown?
No. Exposure usually falls after estimated volatility rises, but gaps, estimation lag, leverage, correlation changes and fast reversals can still produce large losses.
How often should exposure be rebalanced?
There is no universal frequency. Compare fixed schedules and threshold-based rebalancing after realistic costs, using only information available at each decision time.
Primary sources
- Moreira & Muir, Volatility Managed Portfolios, NBER working-paper record and Journal of Finance publication details.
- Harvey et al., The Impact of Volatility Targeting, SSRN paper record.
Educational research only. This article does not recommend an instrument, target or leverage level.