Volatility targeting: scaling exposure to hold risk steady

On this page

Key takeaways

  • Volatility targeting sizes the position from recent swings: it holds more after calm stretches and less after turbulent ones, usually with a cap on borrowing. The rest sits in cash.
  • Because volatility clusters, a turbulent recent stretch tends to be followed by more turbulence, so cutting exposure after one lowers the swings that follow; realised volatility ends up closer to the target than an unmanaged portfolio's.
  • Moreira and Muir (2017) found that scaling factor portfolios down after turbulent stretches raised their Sharpe ratios; later studies found the gain smaller and less reliable on data they had not seen.
  • On MarketHeist's monthly S&P 500 data, a target below the index's own volatility cut its volatility and deepest decline, and its growth rate, by more than it changed the Sharpe ratio.
  • MarketHeist's lab measures recent volatility from past months only, resets the position monthly, and assumes cash earns nothing, borrowing costs nothing and trading is free; those assumptions flatter leverage and are stated beside every result.

Definition

Volatility targeting is a rule that sets how much of a portfolio to hold each period as a target volatility divided by a recent estimate of the portfolio's volatility, so exposure falls after turbulent stretches and rises, possibly above fully invested, after calm ones. The aim is a steadier level of risk over time rather than a fixed allocation.

Also known as: volatility-managed portfolio, volatility scaling, target volatility strategy, risk targeting, constant volatility, vol targeting, volatility control.

How it works

Pick a target volatility. At the start of each month, measure how volatile the portfolio has been over the past twelve months, using only months already finished. If it was twice the target, hold half the portfolio and leave half in cash; if it was four-fifths of the target, hold one and a quarter times the portfolio, borrowing the extra quarter, unless that exceeds the cap on leverage. Next month, measure again and resize.

The rule works through volatility clustering: calm months tend to follow calm months and wild months follow wild ones. Recent volatility is therefore a usable forecast of next month's, and scaling by its inverse keeps realised risk closer to the target than a fixed allocation does. It is like a driver who slows down when the road has been bumpy and speeds up on a smooth stretch: the ride evens out even though the road does not change.

Whether it also improves return per unit of risk is a separate question. If expected returns did not rise in turbulent periods as much as volatility does, holding less during them gives up little return for a lot of risk, and the Sharpe ratio rises. That is the finding of Moreira and Muir. But in a crash followed by a fast rebound, the rule is underweight for the rebound, and a trailing estimate reacts late to a sudden shock. How the two effects net out depends on the asset, the period, the lookback and the frequency of rebalancing.

The target sets the average level of exposure. A target below the asset's long-run volatility keeps the portfolio partly in cash on average, so it grows more slowly in a rising market; a target above it uses leverage on average, which raises growth and losses together and brings the extra drag of volatility described in the leverage guide.

Formula

Exposure in month t
et=min(Lmax,σ*σ^t),σ^t=12·sd(rt−k,…,rt−1)
eₜ
exposure in month t, as a multiple of the portfolio's value (1 = fully invested)
σ*
the volatility target, annualized
σ̂ₜ
the trailing estimate of annualized volatility, from the k months before month t
sd
sample standard deviation (divides by k − 1)
rₜ
the portfolio's monthly total return in month t
k
lookback in months (12 by default)
L_max
the cap on exposure (1.5 by default)
The managed portfolio's return
rtvt=etrt+(1−et)ct,ct=0in this lab
rₜᵛᵗ
return of the volatility-targeted portfolio in month t
cₜ
the cash rate earned on the uninvested part, or paid on the borrowed part when eₜ is above 1

With cₜ = 0 cash earns nothing and borrowing is free. A positive cash rate would add to the return when exposure is below 1 and subtract from it when exposure is above 1.

Moreira and Muir's scaling
ft+1σ=cσ^t2ft+1
fₜ₊₁
the factor portfolio's excess return next month
σ̂ₜ²
the realized variance (uncertainty) of its daily returns over the previous month
c
a constant that gives the managed portfolio the same full-sample volatility as the unmanaged one

The paper scales by inverse variance, not inverse volatility, uses daily data within each month, and sets c with the whole sample, which is not known in real time. MarketHeist's version scales by inverse volatility from monthly returns, with a fixed target and only past data.

Worked example

The S&P 500 (SPY, extended back with its index mutual fund), from Feb 1988 to Sep 2026, held fully invested or resized each month to a volatility target of 10%, measured over the previous 12 months, with exposure capped at 1.5 times.

The target sits below the index's own volatility, so the rule was usually partly in cash: exposure averaged 84%, ranged from 33% to 150% of the portfolio, and sat at the cap in 8% of months, during calm stretches.

Risk came down. Volatility fell from 14.5% to 11.1%, close to the target, and the deepest decline from -50.8% to -38.8%.

Growth came down too, from 11.3% to 9.0% a year, largely because the portfolio held less than all of the index on average. Per unit of volatility the two were close: a Sharpe ratio of 0.84 against 0.81 at a risk-free rate of 0.

So on monthly data with a one-year lookback, targeting volatility on the S&P 500 mainly changed the shape of the ride, not the reward per unit of risk. The larger Sharpe gains in the research come from other assets, daily data and full-sample scaling; the lab lets you shorten the lookback, raise the target or the cap, and switch to a stocks-and-bonds portfolio to see how much each choice matters.

Data through Sep 2026. Reproduce id 582599f6115a1fa4 (this guide, its lab setup and the data month).

Interactive lab

Growth of $10,000 fully invested and volatility-targeted, with the monthly exposure below, and the table of growth, volatility, deepest decline, Sharpe ratio and exposure. Move the target, the lookback or the leverage cap, or switch portfolios.

Growth of $10,000, plain and volatility-targeted at 10%: S&P 500 (100%)S&P 500 (100%), Feb 1988 to Sep 2026. Each month the book holds 10% divided by the volatility of the previous 12 months, capped at 150%; what is not invested earns 0 and an exposure above 100% borrows at 0. $10,000 grows to $633,728 plain and $284,933 targeted.$20k$50k$100k$200k$500k19881993199820032008201320182023PlainTargeted at 10%
Growth of $10,000 on a log scale. Feb 1988 to Sep 2026.
Exposure of the volatility-targeted book: S&P 500 (100%)Share of the book held each month, Feb 1988 to Sep 2026. Average 84%, highest 150%, lowest 33%; at the 150% cap in 8% of months.0%20%40%60%80%100%120%140%19881993199820032008201320182023100%Exposure
Share of the book held each month; the dashed line is 100%. Feb 1988 to Sep 2026.
MeasurePlainTargeted at 10%
CAGR11.3%9.0%
Volatility (annualized)14.5%11.1%
Maximum drawdown-50.8%-38.8%
Sharpe ratio0.810.84
Growth of $10,000$633,728$284,933
Average exposure100%84%
Highest and lowest exposure100%150% and 33%
Months at the 150% capnone8%

Exposure is the target over the trailing volatility (the sample standard deviation of the 12 months before each month, times the square root of 12), capped at 150%. Money not invested earns 0, and an exposure above 100% is borrowed at 0. Risk-free rate 0. Feb 1988 to Sep 2026.

How MarketHeist computes it

Every choice an implementation has to make, as the code makes it today. The page is checked against the code each time it is built.

ChoiceHow it is computedIn the code
DataMonthly total returns (dividends reinvested). Funds younger than the period shown are extended backward with older funds, spliced on returns, not prices.
WindowThe worked example and the lab compare the plain portfolio and the targeted one over the same Feb 1988 to Sep 2026 (464 months), which starts once the longest lookback the lab offers (36 months) has history, so the lookback slider does not move the window.
Portfolio streamsA fund is its proxy-extended monthly return stream. A built-in portfolio is read as the MarketHeist MCP reads it: its composed stream with its own rebalancing rule and overlay. A custom mix is rebalanced yearly.
Volatility targetingEach month the exposure is the target divided by the trailing volatility, capped at the leverage limit. The trailing volatility is the sample standard deviation of the monthly returns of the lookback months before the month (never the month itself), times the square root of 12. The portfolio earns the exposure times that month's return; the part not invested earns 0, and an exposure above 100% borrows at 0, so no financing cost is charged. The plain comparison is the same stream fully invested.
ExposureEach month, min(cap, target ÷ trailing volatility). The trailing volatility is the sample standard deviation (divides by N − 1) of the lookback months strictly before the month, times √12, so a month's own return never sets its exposure.volTargetSidecontent-engine/lab/compute4.ts
Cash, borrowing and costsThe uninvested part earns 0, an exposure above 1 borrows at 0, and resizing is free: no interest, spreads, commissions or taxes. A real version would earn interest on cash, pay interest on borrowing and pay for monthly trades.volTargetSidecontent-engine/lab/compute4.ts
WindowBoth columns cover the same months, starting after 36 months of history whatever the lookback, so moving the lookback slider does not move the window.VT_ANCHOR_MONTHScontent-engine/lab/compute4.ts
RangesTarget from 6% to 20% a year, lookback from 3 to 36 months, leverage cap from 1 (never above fully invested) to 3.VT_LEVERAGE_MAXcontent-engine/lab/compute4.ts
MetricsCAGR, volatility (sample standard deviation of monthly returns times √12), deepest decline and Sharpe ratio (risk-free rate 0) as everywhere on MarketHeist. Months at the cap are those whose uncapped exposure reached the cap.metricsFromReturnspackages/portfolio/engine/portfolio.ts
Backtester's target-volatility leverageDifferent: the backtester computes ONE multiplier for the whole backtest, target ÷ the strategy's realised volatility over the entire period, and applies it to every bar. That uses the full history, including the future relative to each bar, and does not vary over time; this lab's rule resizes every month from past data only.resolveLeveragepackages/shared/engine/backtest.ts
Engine functionsThe functions the portfolio app and the MarketHeist MCP tools run, called here on the same monthly returns.nanstdpackages/shared/engine/utils.tsmetricsFromReturnspackages/portfolio/engine/portfolio.ts

Volatility targeting compared with its neighbours

Volatility targeting and related ways to manage risk
ApproachWhat sets exposureReacts toTypical cost
Fixed allocationNothing; always the sameNothingRisk swings with the market
Volatility targetingTarget ÷ recent volatilityRecent turbulenceUnderweight in rebounds; trading; borrowing above full exposure
Trend filterPrice against its moving averageFalling pricesWhipsaws and missed rebounds
Risk parity across holdingsEach holding's volatility and correlationsRelative risk of the holdingsOften needs leverage to reach a return goal
Constant leverageA fixed multipleNothingLarger drawdowns and volatility drag

Limitations

  • Cash earns nothing and borrowing is free in the lab. With interest on cash the low-exposure months would earn a little more; with a borrowing rate, every month above full exposure would cost more, which can erase the benefit of a high target.
  • Monthly rebalancing and a trailing monthly estimate react slowly. A crash that happens within a month is taken at whatever exposure was set before it, and the rule cuts exposure after the damage, often just before a rebound.
  • Results depend on the target, the lookback and the cap; a setting picked because it looked good on this history is subject to the same overfitting as any other.
  • Trading costs and taxes from monthly resizing are left out.
  • One historical path, with long histories extended backward with older funds or indexes.

Frequently asked questions

Does volatility targeting improve returns?
It usually lowers volatility and the deepest decline. Its effect on return per unit of risk is mixed: Moreira and Muir (2017) found higher Sharpe ratios for several factor portfolios, while later work found smaller and less reliable gains out of sample, and on monthly S&P 500 data the change was small.
Why does volatility targeting work at all?
Because volatility clusters: recent volatility is a fair forecast of near-term volatility, while expected returns have not risen as much in volatile periods. Holding less when volatility is high gives up relatively little return for a large cut in risk.
What lookback do volatility-targeting strategies use?
Anything from a few weeks of daily returns to a year of monthly ones. Shorter lookbacks react faster but are noisier and trade more; longer ones are steadier and slower. The lab offers 3 to 36 months of monthly returns.
Is volatility targeting the same as risk parity?
No. Volatility targeting scales the whole portfolio up or down over time to hold its total risk near a target. Risk parity sets the weights of the holdings so each contributes equally to risk. The two are often combined: a risk-parity mix scaled to a volatility target.
Does a target above the asset's own volatility make sense?
It means borrowing on average, which raises growth and losses together and adds volatility drag. In the lab borrowing is free, which flatters high targets; the leverage guide shows the effect.

References

  1. Fleming, J., Kirby, C. and Ostdiek, B. (2001) The Economic Value of Volatility Timing. Journal of Finance 56(1).
  2. Moreira, A. and Muir, T. (2017) Volatility-Managed Portfolios. Journal of Finance 72(4).
  3. Harvey, C. R., Hoyle, E., Korgaonkar, R., Rattray, S., Sargaison, M. and Van Hemert, O. (2018) The Impact of Volatility Targeting. Journal of Portfolio Management 45(1).
  4. Cederburg, S., O'Doherty, M. S., Wang, F. and Yan, X. (2020) On the Performance of Volatility-Managed Portfolios. Journal of Financial Economics 138(1). Out-of-sample evidence.
  5. Barroso, P. and Santa-Clara, P. (2015) Momentum Has Its Moments. Journal of Financial Economics 116(1).

Try it on your own holdings

Set target-volatility leverage on a backtest →

Or ask an assistant connected to the MarketHeist MCP server (https://marketheist.io/mcp), which computes the same figures:

Using MarketHeist.io's monthly S&P 500 history, what happens if exposure is resized each month to a 10% volatility target using the previous 12 months, capped at 1.5x with cash at 0%: how do CAGR, volatility, max drawdown and Sharpe compare with holding the index, and how often is the cap hit?

Ask an AI assistant

The prompt opens in a new tab, already filled in. An assistant without the MarketHeist connector answers from its own knowledge, not from the figures on this page.

Historical simulation, not advice. Computed by the MarketHeist portfolio engine from monthly total returns, data through Sep 2026. Markdown version of this page: volatility-targeting.md.