What Is a Conditional Variance Swap?
Conditional variance swap pays on realised variance accumulated only during periods when the underlying price sits within a specified range, ignoring movement outside it. It lets a position target volatility in one part of the price distribution rather than across all of it.
A variance swap pays on how much a price moved. A conditional variance swap pays on how much it moved while it was somewhere specific — and that restriction is the entire product.
How it works
A price range is specified in the contract. Realised variance is accumulated day by day, but only on days when the underlying sits inside that range.
Days outside contribute nothing. However violently the price moves above or below the barrier, those moves do not enter the calculation.
Which separates volatility by location. A downside conditional swap counts only movement below a level, so it is a position on how turbulent declines are, specifically.
Why anybody wants this
Volatility is not uniform across the price range. Markets move differently when falling than when rising, and a plain variance swap averages those two behaviours into one number.
A dealer hedging options has region-specific exposure. Their book’s sensitivity is concentrated around particular strikes, and a hedge that pays on movement everywhere is a poor match for it.
So the conditional version is a precision tool. It pays on exactly the region where the exposure lives and costs nothing for the regions where it does not.
A worked example
A downside conditional variance swap with a barrier at 90, on something currently trading at 100.
For as long as the price stays above 90, nothing accumulates. The buyer pays their fixed strike and receives nothing, and the contract is inert.
The price falls to 85 and becomes turbulent. Now every day’s movement counts, and turbulent declines produce large variance — so the contract pays substantially.
Which is the shape somebody wanting crash protection actually needs. A plain variance swap would have been accumulating small amounts in the calm 100-to-110 region all along and charging for it.
Time in range is the hidden variable
The payoff has two components. How volatile the price was inside the range, and how many days it spent there at all.
A contract whose range is never entered pays nothing regardless of how right the volatility view was. The direction call and the volatility call are both required.
And the accumulated variance is usually divided by the days in range, or by the total days, depending on the contract — and those two conventions produce very different payoffs.
Which is why these are institutional instruments. The economics depend on a convention buried in the documentation, and an otherwise identical trade under the other convention is a different position. Anything with that property is a professional product, and describing it as anything else would be misleading.
The original data
On this site’s shared series the median bar range is 0.493, the ninetieth percentile 1.101 and the largest single bar 2.338. ATR14 has a median of 0.5994 and a ninetieth percentile of 0.7954.
That gap between median and ninetieth percentile is what a conditional swap is trying to isolate. The large bars are not spread evenly across the price range; they cluster in declines, and a contract counting only that region captures them without paying for the calm ones.
And the drawdown measurement gives the frequency: 95% of bars sit below a prior peak, with a maximum decline of 3.76% and a longest below-peak stretch of 73 bars, which finished +3.61%.
What it tells you even if you never trade one
Volatility has a location. The same asset is a different instrument on the way down than on the way up, and pricing reflects that.
Which is what the volatility smile is. Implied volatility differing by strike is the market quoting different volatilities for different regions of the price range — precisely the thing a conditional variance swap isolates and trades.
And it explains why downside protection is expensive. Not because sellers are pessimistic, but because measured downside volatility genuinely exceeds measured upside volatility, and the price reflects the measurement.
That reframing is worth more to most readers than the instrument is. “Puts are overpriced” is a common complaint and usually wrong — the asymmetry being charged for is real and observable in the data.
How it is priced and hedged
A variance swap can be replicated with a strip of options across every strike, which is what makes it priceable at all — the fair strike is read off the option surface rather than forecast.
The conditional version narrows that strip. Only options in the relevant region contribute, so its price comes from one part of the volatility surface instead of all of it.
Which is why it is cheaper than the unconditional version and why the two are not substitutes. You are buying a smaller piece of the same surface, and the discount is exactly the value of the piece you did not buy.
And the hedge is dynamic. A dealer holding one must adjust as the price approaches the barrier, since the contract’s sensitivity changes sharply there — which makes the barrier region itself a place where hedging flow can affect the price.
When it fails
The characteristic failure is being right about volatility and wrong about location. A buyer purchases downside conditional variance expecting turbulence, and the turbulence arrives as a violent rally instead. The market moved enormously, the volatility view was correct, and the contract pays nothing because the price never entered the range. Both conditions had to hold and only one did — which is the structural cost of the precision that makes the instrument cheaper than an unconditional one.
A second failure is a gap straight through the range. A price that jumps from above the barrier to well below it accumulates almost no days inside, and a crash can produce a small payoff.
A third is misreading the averaging convention, which changes the payoff materially.
A fourth is treating it as cheaper protection. It is narrower protection, priced accordingly.
And a fifth is assuming volatility is one number. It is not, and this instrument exists because it is not.
Related
Volatility risk covers the general exposure this instrument slices up. Volatility covers the measure itself. And volatility smile covers the same region-dependence as it appears in option pricing.
This is a specialist instrument and it teaches something general: volatility is not one number. It is different in different parts of the price range, and once you accept that, the whole apparatus of the volatility smile stops looking like a market quirk and starts looking like a measurement.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.