Volatility targeting is a position-sizing rule. It holds a portfolio's risk near a fixed level by cutting exposure when volatility rises and adding exposure when it falls. The trader sets a risk target, measures recent volatility, and sizes the position so that dollar risk stays roughly constant. This guide covers the formula, worked examples, the lookback choices that change results, what the research shows, the costs, and how the idea applies to gold (XAUUSD).
TL;DR
- Volatility targeting sets position size as target volatility divided by expected volatility. High volatility means a smaller position. Low volatility means a larger one.
- Academic work finds the clearest Sharpe ratio gains in equities and credit. For commodities and other assets, the main effect is fewer extreme returns.
- Lookback window, forecast method and rebalancing speed decide turnover and trading costs. Costs can erase the benefit if ignored.
- The rule reacts after volatility has already risen. It manages risk, but it does not promise a smaller loss.
- In gold, set a leverage cap before anything else. Gold can move fast, and leverage multiplies both sides.
Table of Contents
- What volatility targeting does
- The formula
- Worked examples
- Lookback, forecast and rebalancing: the choices that matter
- What the research shows
- Risks and costs
- Applying it to gold (XAUUSD)
- Our perspective
- Copy a gold strategy instead of building one
- FAQ
- Sources
What volatility targeting does
Most traders size positions by conviction or by a fixed amount. Volatility targeting sizes them by risk. The aim is that a $10,000 position in a quiet market and a smaller position in a stormy market carry about the same expected daily swing.
The rule works because volatility clusters. Calm periods tend to follow calm periods, and turbulent periods follow turbulent ones. Campbell Harvey and coauthors explain this as the basis of the approach. Tomorrow's return is very hard to forecast. Tomorrow's volatility is much easier to forecast. Volatility targeting uses that forecasting edge for the one decision it can inform: how big the position should be.
The approach is common in portfolio management, in systematic futures trading and in risk-managed index products. It also works for a single instrument.
The formula
The core rule for a single asset is simple. Scale exposure by the ratio of the target volatility to the expected volatility:
Exposure = target volatility / forecast volatility
When the ratio is above 1, the portfolio needs leverage to hit the target. When it is below 1, the portfolio holds less than full exposure and the rest sits in cash.
Most practical systems add two guardrails:
- A leverage cap, so a very quiet market cannot push exposure to extreme levels.
- A floor, usually zero for long-only accounts.
Worked examples
The examples below use round numbers to show the mechanics. They are illustrations, not trade signals.
Example 1: one asset, 10% target.
| Realized annual volatility | Exposure formula | Exposure | Result |
|---|---|---|---|
| 20% | 10% / 20% | 0.50x | Half the capital invested, half in cash |
| 10% | 10% / 10% | 1.00x | Fully invested |
| 5% | 10% / 5% | 2.00x | Leveraged 2:1, or capped lower |
Example 2: a $100,000 portfolio. With a 10% target, realized volatility of 8% gives a multiplier of 1.25, which lifts exposure to $125,000. If volatility jumps to 15%, the multiplier falls to about 0.667, and exposure shrinks to about $66,700. The portfolio's risk lands back near 10% either way.
Example 3: gold with a cap. Take a 15% annual target for gold and a leverage cap of 2.0. If gold's forecast volatility is 12%, exposure is 1.25x. If it is 25%, exposure is 0.6x. If it falls to 6%, the formula says 2.5x, but the cap holds it at 2.0x.
The daily version is what most systems run. Divide the annual target by the square root of 252 trading days. A 15% annual target becomes roughly 0.95% per day.
Lookback, forecast and rebalancing: the choices that matter
The formula is the easy part. The forecast of volatility is where strategies differ.
Lookback window. A short window reacts fast but jumps around. A long window is smoother but slower. A 20-day rolling standard deviation has a known flaw: one large shock spikes the estimate, then it drops out of the window about 21 days later and exposure jumps up in one step. Exponentially weighted averages avoid that cliff, because old observations fade gradually instead of dropping out.
Forecast method. Dimensional's S&P 500 case study (data 1997 to 2023) found that simple forecasts with few parameters can stabilize volatility as well as complex models. Combining implied volatility with simple estimates from past returns gave good control with lower turnover.
Rebalancing speed. The same study found that daily turnover was much higher for forecasts built on very short windows. Averaging over about 21 days cut it sharply. Slower rebalancing, or a band around the target that triggers trades only on larger deviations, keeps trading costs in check.
A practical starting point for a trader building a rule:
- Pick an annual risk target you can live through in a bad month.
- Estimate volatility with an exponentially weighted average over roughly 10 to 30 trading days.
- Convert the annual target to daily by dividing by the square root of 252.
- Cap leverage.
- Rebalance on a schedule or when exposure drifts beyond a set band.
- Test with realistic costs before risking capital.
What the research shows
The evidence is real but narrower than the marketing around it.
Equities. Moreira and Muir found that volatility-managed portfolios on the market produced an alpha of 4.9% and raised the buy-and-hold Sharpe ratio by about 25% in US data from 1926 to 2015.
Many asset classes. Harvey and coauthors tested more than 60 assets with daily data going back as far as 1926. Sharpe ratios rose with volatility scaling for equities and credit. The effect on bonds, currencies and commodities was negligible. Across all of them, scaling consistently reduced the likelihood of extreme returns.
The CFA Institute summary of that paper ties the equity result to the leverage effect. Falling prices raise volatility, volatility scaling cuts positions, and the rule behaves like a mild momentum overlay.
The practical reading for a gold trader is that the main, well-documented benefit is steadier risk and fewer extreme outcomes. A higher Sharpe ratio is not guaranteed. That is a sound reason to use the rule: it is a risk tool first.
Limits of the evidence. Bongaerts and coauthors argue that some earlier results rely on scaling factors built with hindsight, and that a conventional version can overshoot its target, which raises maximum drawdowns and tail risk. Macrosynergy also notes that mechanical volatility targeting can contribute to herding when many investors cut risk at the same time.
Risks and costs
Volatility targeting seeks to keep risk steady. It does not promise a smaller loss. Know these points before you use it.
- It reacts after the fact. Volatility is estimated from the past, so exposure falls only after volatility has risen. A sudden gap can hit a full-size position.
- V-shaped recoveries. Exposure stays low while realized volatility is still high, so part of a quick rebound can be missed.
- Calm before a shock. In a quiet market the rule may be at its leverage cap just before volatility jumps. This is why the cap matters.
- Trading costs. Bianchi's study of smoothing found that for a plain realized-variance method, the average Sharpe ratio fell from 0.23 to 0.17 after 14 basis points of costs and turned negative at 50 basis points. A smoothed forecast held up far better. Smooth the forecast and limit turnover.
- Leverage. Above 1.0x, every loss is magnified. Read the broker's margin and stop-out rules, and size the cap so a bad week cannot force a close.
Applying it to gold (XAUUSD)
Gold suits a risk-based sizing rule because its volatility changes a lot across regimes and across the trading day. Our guide to XAUUSD volatility by time of day draws on 23 years of intraday data and shows how much movement concentrates in certain hours. A fixed lot size means very different risk in a quiet Asian session and in the London and New York overlap.
A recent academic example comes from Singha and coauthors, who tested volatility targeting on gold futures from 2015 to 2025. They targeted 15% annualized volatility with a leverage cap of 2.0, using an exponentially weighted variance forecast. Their out-of-sample results are reported net of costs, with a Sharpe ratio of 2.88. These are one paper's reported figures from a specific model and period. They are not a forecast for any account.
What a gold trader should set before using any such rule:
- The risk target. Choose it from the drawdown you can tolerate. Our explainer on maximum drawdown for gold copy traders shows why a 20% loss needs a 25% gain to recover.
- The leverage cap. Read how gold traders use 2x exposure with care and the leverage risks specific to XAUUSD.
- The stop logic. ATR-based stops fit naturally with volatility sizing, because both use the same measure of market movement. See our XAUUSD volatility strategy notes.
- The event calendar. Scheduled macro releases create gaps that a backward-looking estimate cannot see. Reduce size ahead of them by rule, not by feel.
- The trading hours. Decide whether the volatility estimate uses all hours or only the sessions you trade.
Gold trading is high risk. Leveraged positions can lose more than expected in fast markets. Anyone sizing real capital should speak with a licensed financial adviser about their own situation.
Our perspective
We look at gold through one lens: control the risk first, and the returns have a better chance of following. Volatility targeting is one way to put that into practice. Sonic AI is another.
Sonic AI trades only gold (XAUUSD), with an algorithm-driven strategy run by an experienced team. The strategy aims for an 80% win rate, and the results are verified on MyFXBook. They show 18 consecutive winning months and drawdown under 1%. We state these as our own claims, and the live record is public so you can read it for yourself.
Most people who want gold exposure do not want to estimate volatility, set caps and rebalance every day. Copy trading through COPYX removes that work. Trades are copied automatically, so no trading knowledge is needed, and account size does not change the process.
The lesson from the research is the lesson we build around: limit the size of your worst day, and you keep the ability to compound. Our write-up on how a low drawdown strategy works explains the approach.
Paulo
Copy a gold strategy instead of building one
Building a volatility-targeted system takes data, code, testing and daily discipline. Readers who want gold exposure without that workload can look at Sonic AI, an auto-trading strategy for XAUUSD that copies trades to the client's account through COPYX.
The live results are on the Sonic AI MyFXBook results page, and the Sonic AI home page explains setup. Fees are explained on the performance fee page. Amplified account options are described on the 12X and 24X amplified account page.
Before copying any strategy, set your own limits. Our guides on copy trading parameters and copy trading drawdown show how to choose caps and stop levels that fit your account.
FAQ
Is volatility targeting the same as position sizing?
It is a type of position sizing. Standard sizing might risk a fixed percentage per trade. Volatility targeting adjusts size continuously so that the portfolio's expected volatility stays near a set level. The two can be combined, with volatility setting the size and a stop defining the risk per trade.
What volatility target should I choose?
Choose the target from the loss you can accept, not from the return you want. Many examples use 10% to 15% annualized. A higher target means larger swings and larger drawdowns. Test the target on past data, including the worst months, and check that the account could survive them with your leverage cap in place.
Does volatility targeting work for gold?
The broad research finds the clearest Sharpe gains in equities and credit. For commodities the effect on Sharpe is small, but scaling still reduces extreme returns. For gold, that makes it a risk-control method more than a return booster. A gold futures study reports strong results with a 15% target and a 2.0 leverage cap, but those are one model's results.
How often should I rebalance?
Only as often as costs allow. Daily rebalancing on a very short lookback produces high turnover. A smoothed forecast, a longer window or a tolerance band around the target keeps trading down. Test with realistic spreads and slippage for the instrument you trade.
Sources
- An Introduction to Volatility Targeting (QuantPedia)
- The Impact of Volatility Targeting (Man Group)
- Forecasting and Managing Volatility: An S&P 500 Case Study (Dimensional)
- Volatility-Managed Portfolios (Moreira and Muir)
- The Impact of Volatility Targeting (Harvey et al.)
- The Impact of Volatility Targeting, digest summary (CFA Institute)
- Conditional Volatility Targeting (Bongaerts et al.)
- Equity alpha through volatility targeting (Macrosynergy)
- Smoothing volatility targeting (Bianchi)
- Forecast-to-Fill (Singha et al.)