Volatility Futures: Trade the Contract, Understand the Curve

A view that an equity volatility index will fall is not yet a trade. The investor must identify an instrument, its settlement definition, maturity, financing and sensitivity to the quoted index. A volatility future, a rolling futures product, an equity straddle and a put hedge can respond very differently to the same change in the displayed volatility level.

This reference report merges the volatility-futures lessons into one account of instrument selection, term-structure interpretation and controlled mean-reversion research. The source collection contains conflicting statements about the availability of particular domestic contracts. Accordingly, the mechanics below are stated conditionally for a market where the relevant contract exists; this report does not assert current contract availability or present a historical strategy as immediately executable.

Spot volatility is a calculation; a future is a contract

A volatility index derived from an option strip summarises a specified horizon of option-implied variance in annualised volatility units. It is a calculated reference value, not a security that can necessarily be bought at the displayed number.

A future has a quoted price today and settles against a defined reference value on a future date. Its terminal reference may itself represent volatility over a further option horizon. The investor therefore needs to know both the futures expiry and what the settlement value measures. “One-month volatility” and “a future expiring in one month” need not refer to the same time interval.

This is why comparing a future directly with today’s spot volatility can be misleading. The difference can reflect expected mean reversion, the distribution of future variance, risk premia and the contract’s settlement convention. It is not simply an ordinary cash-and-carry basis that can be arbitraged by buying the spot index.

Read the curve as a conditional distribution of future risk

An upward-sloping futures curve is commonly called contango; a downward-sloping one is backwardation. In a calm regime, later volatility can be priced above the current level because some normalisation is expected. During acute stress, near-term volatility can exceed later volatility because the market prices a possible decline after the shock.

The source describes contango on roughly 80% of days and backwardation on roughly 20%. Those are retained historical teaching claims, not verified frequencies for every market or period. The curve shape can remain unusual for longer than a strategy’s financing capacity.

The key analytical questions are more specific than the label. Which interval is expensive relative to the others? Does the apparent event premium lie before, inside or after the futures settlement window? How much of the view is on the level of volatility and how much is on a change in curve shape?

Figure 01 · Qualitative comparison

One volatility label can hide several different payoffs

One volatility label can hide several different payoffs

Object	What determines the value or payoff	Key distinction
Volatility index	What determines the value or payoffA specified option-price calculation	Key distinctionA displayed measurement is not automatically investable
Volatility future	What determines the value or payoffA defined index settlement at its contract expiry	Key distinctionExposure is to future settlement, not today’s spot index
Variance swap	What determines the value or payoffRealised variance over an observation interval	Key distinctionPays on squared volatility units
Equity option or straddle	What determines the value or payoffUnderlying payoff, implied repricing and hedge path	Key distinctionIncludes gamma, theta and potentially directional exposure

Convergence is conditional on the realised settlement

A retained teaching example places a six-month futures price at 16.7 and current spot volatility at 13.2. If the future eventually settles at 13.2, a long position loses 3.5 volatility points before costs. The numerical difference is clear:

$$13.2-16.7=-3.5.$$

The loss does not occur merely because time passes. It depends on the settlement outcome. If the future settles at 25 instead, the long position gains 8.3 points. A futures price converges to its own settlement reference at expiry; it is not guaranteed to converge to the spot value observed six months earlier.

Multiplying the point change by the contract multiplier and number of contracts gives cash profit. Dividing by initial margin produces a return on margin, not a return on fully funded economic exposure. This denominator can make both gains and losses appear very large and should be displayed separately from the cash outcome.

The reverse logic applies in a stressed downward-sloping curve. A short future can benefit if a temporary spike fades, but further escalation can produce large adverse marks before any eventual mean reversion. The shape of the curve does not finance those marks.

Rolling exposure adds a second mechanism

A constant-maturity product continually changes the mix of futures it holds. As the front contract approaches expiry, the portfolio sells some front exposure and buys later exposure. Its return is determined by the change in the held futures, the changing weights, fees and any leverage or daily-reset mechanism.

The shorthand “selling low and buying high” describes the direction of a roll in an upward-sloping curve, but it is not a complete self-financing profit calculation. A proper attribution measures the performance of each contract over the holding interval and then the contribution of moving the maturity exposure.

The source cites long-volatility rolling products losing about 99.9% over a particular decade despite a broadly similar beginning and ending spot-volatility level. It also cites a 68% loss in one calm year. Those are retained source-reported product comparisons, not independently verified returns. They illustrate why the path of the futures curve, fees and product design must be studied instead of treating the spot index chart as the investment’s return series.

Leverage and daily reset introduce further path dependence. The cumulative return of an inverse daily-reset product is not simply minus the cumulative return of its underlying reference. Large alternating daily changes can erode value even when the reference ends near its starting level. Termination and acceleration provisions can also matter in extreme moves.

Figure 02 · Conceptual framework

Convergence is to eventual settlement, not today’s index

How the parts connect

1. AFutures entry pricestarting priceTrade outcome
2. BCurrent spot indexcontext, not a fixed endpointFuture market path
3. CFuture market pathleads to eventual observationDefined settlement
4. DDefined settlementterminal referenceTrade outcome

Futures, variance and options are related but not equivalent

Related instruments transfer different volatility risks

The square-root relationship between variance and volatility also introduces convexity. The expected square root of future variance is generally not the square root of expected variance. A volatility future is therefore not mechanically identical to the square root of a forward variance-swap strike.

Calling a straddle a “synthetic volatility future” can be useful only as a loose exposure analogy. The investor inherits an option payoff and a hedging problem, not the exact futures settlement. Any comparison should identify the approximation and quantify its failure in scenarios where spot or the surface changes substantially.

Mean-reversion signals need more than an extreme z-score

The source proposes measuring a volatility level relative to its trailing mean and standard deviation:

$$z_t=\frac{V_t-\bar V_t}{s_{V,t}}.$$

Here $V_t$ is the chosen volatility observation, $\bar V_t$ its trailing mean and $s_{V,t}$ the trailing standard deviation. Suggested thresholds include +2 for a high reading and −1 for a low reading. They are candidate rules to test, not probabilities of profit.

A high z-score may represent a temporary dislocation or a new regime. The lookback itself can be misleading when the preceding period was unusually calm. A curve confirmation, an event classification or a persistence estimate can add information, but each added filter introduces another parameter and another opportunity for overfitting.

Illustrative historical-style recovery windows

Retain these as illustrative classifications. Events do not arrive with a reliable duration label, and classifying them using their eventual recovery date would give a backtest knowledge unavailable at entry.

Figure 03 · Conceptual framework

The return of a rolling exposure has several moving parts

How the parts connect

1. ACurrent futures holdingsmarket exposurePrice changes
2. BCurrent futures holdingsscheduled transitionRoll transaction
3. CPrice changesmark-to-market effectRealised portfolio return
4. DRoll transactionchanges future holdingsRealised portfolio return
5. EProduct mechanicsimplementation effectRealised portfolio return

Preserve a historical call without pretending it was easy

One historical lesson describes volatility falling from 26.8 to 15.5 around an event-resolution period, a decline of approximately 42.2%. Different companion documents place parts of the rise and fall on inconsistent dates. The arithmetic can be retained while the unresolved chronology prevents treating the example as an audited executable trade.

The original trading interpretation was to sell an elevated event premium in anticipation of normalisation. That is the historical call. Recording it faithfully also requires acknowledging the alternative path: the underlying could gap adversely and volatility could rise further before falling. An attractive beginning-to-end index change does not show whether the position survived the intervening margin calls or hedge losses.

The example also quotes a sold straddle premium of 460 and a later value of 110, implying 350 points of gross option-value decline. Actual profit would require the exact legs, quantity, timestamps, hedge trades and costs. It cannot be inferred from the option-price difference alone if the position was dynamically hedged.

Size in cash loss and settlement exposure

For a futures trade, define the cash effect of a one-point move and then apply adverse scenarios. A volatility-level rise of 30% and a rise of 30 points are very different stresses. A position entered at 20 reaches 26 under the first scenario and 50 under the second.

The teaching source suggests a per-trade vega exposure cap equivalent to 0.5% of portfolio equity, but a future’s point sensitivity and an equity option’s vega are not interchangeable without a mapping. State the instrument’s actual cash-per-point exposure and preserve any option approximation separately.

Use both level and curve shocks for a spread. A front-versus-back trade can have a modest parallel-level sensitivity and still suffer if the front maturity jumps relative to the back. Margin models can also change as volatility increases, raising cash requirements even if the economic thesis remains unchanged.

Stops should be part of a plan but not assumed to execute at the trigger. A gap, an illiquid quote or a market interruption can move the realised exit materially beyond the intended level. The scenario budget should therefore include delayed execution and a stress spread, not just a perfectly filled stop.

Settlement basis can dominate the final holding interval

A displayed volatility index and the settlement calculation of a derivative can differ because of the option observations, timestamps and contract rules used. Near futures expiry, this settlement basis deserves direct attention. A strategy that treats the last displayed index value as the guaranteed settlement can mismeasure the final profit even if it forecasts the broad direction correctly.

The sensitivity of a futures price to spot volatility can also change across maturities and regimes. A far-dated contract may react less to a temporary spot spike because it refers to a later volatility window. As that maturity approaches, its relationship to the eventual reference changes. A hedge ratio estimated in calm markets may consequently fail during the event it was meant to offset.

For a spread, record separate shocks to each maturity instead of relying only on a historical correlation. The curve can twist when one expiry contains an event and another does not. A parallel move leaves some spreads nearly unchanged while a localised front-end shock produces a large loss. The risk report should show both possibilities.

This is another reason to distinguish a futures calendar from an equity-options calendar. They may both express a term-structure view, but the futures spread settles on volatility references while the option spread retains strike and spot dependence. An approximate substitution should be justified by scenario behaviour over the intended holding interval, not simply by matching two initial sensitivity numbers.

An investable study needs contract histories, expiries and settlement definitions, rather than a continuous spot-volatility chart alone. It should identify the roll rule, quote used for execution, cost assumptions and the time at which each signal became observable. It must also distinguish available contracts from teaching proxies.

The final report should separate spot-reference movement, futures holding profit, roll contribution, leverage effects and fees. That attribution shows whether the proposed return comes from forecasting a volatility decline, accepting a term premium, taking concentrated short-tail exposure or merely choosing a favourable denominator.

The durable lesson is to identify the trade that actually exists. A sound view on volatility can still lose money in the wrong maturity or product. Contract definition, curve behaviour and funding capacity are part of the investment thesis from the beginning.