How to Build the Volatility Term Structure: Two Methods

One curve, two completely different ways to build it

Most people who say "the term structure is in backwardation" are looking at VIX futures. Fair enough, that is the curve everyone quotes. But there is a second construction, built straight from SPX options, that answers questions the futures curve physically can not. Plenty of traders spend years on the first without knowing the second exists.

Quick grounding first. Implied volatility is forward-looking: it is what the options market collectively prices as the expected magnitude of movement over a given horizon. A single IV number covers one expiration. Line those expectations up across every available horizon and plot them, and you have the volatility term structure.

In a calm market the curve slopes upward. Options expiring further out carry higher IV than near-term options, because there is simply more calendar time for something to go wrong. That is contango: a progressively larger uncertainty premium for longer horizons. It is the default state most days.

Under stress, the relationship inverts. Traders rush to buy immediate protection, short-term IV spikes above long-term IV, and the curve slopes downward.

That is backwardation, and in my experience it is one of the more reliable tells that fear has turned acute rather than chronic.

Illustrative diagrams, not real market data

So far, standard. What matters for this article is that the same curve can be built from two completely different data sources. Both describe the same underlying reality. Each has advantages the other does not, and the choice depends on what you are trying to learn.

Method 1: the VIX futures term structure

Start with what the VIX actually is: a calculated number, nothing tradeable. It represents the options market's expectation of 30-day S&P 500 volatility, derived from a weighted average of SPX options prices across a range of strikes. You can not buy or sell the VIX itself. What you can trade are VIX futures, standardized agreements to exchange cash at a specified future date based on where the VIX settles.

VIX futures expire on specific dates (monthly, plus weekly maturities added since 2015). Each contract has its own price, and plotting those prices against their expiration dates gives you the VIX futures term structure: a snapshot of where the market expects the VIX to be at each future settlement date.

In contango, each successive contract sits above the prior one. The market expects higher volatility in the future than today, and that is the default state, because short volatility is a risk sellers want to be paid for. In backwardation the near-term contracts trade above the back months, meaning the market expects the current elevated volatility to decay over time.

The case for this method is convenience. The data is everywhere: major terminals, CBOE, most brokers publish /VX prices in real time. Each point on the curve is a single futures price, so there is no calculation to do, you just plot and read. And if you trade volatility ETPs like VXX, UVXY, or SVXY, this curve directly dictates their performance. You need it as context regardless of anything else in this article.

The cost of that convenience is resolution. Expirations are weekly and monthly, with nothing in between, so the curve gives you no day-by-day detail. It cannot tell a CPI day apart from the day after. For a macro regime read, that is fine. For event timing, it is a blunt instrument.

Method 2: the SPX options implied volatility term structure

The second method skips futures entirely and goes to the source: the implied volatility embedded in each SPX options expiration. For every available expiration in the chain, you extract the IV and plot it against days-to-expiration. That is the whole idea. A term structure built purely from the options market, no futures data required.

The construction detail that trips people up is which option to use at each expiration. You can not just pick any strike. Options at different strikes carry different IVs because of the volatility skew (out-of-the-money puts typically trade at higher IV than at-the-money options). Mix strikes across expirations and you smear skew distortion into the curve, burying the term structure signal. The fix is to standardize on a single moneyness at every expiration.

The standard choice is at-the-money implied volatility: the IV of the option struck closest to current spot, or the delta-neutral strike where call and put have roughly equal delta magnitudes. Using ATM IV at every expiration isolates the time dimension of implied volatility from the skew dimension. The resulting curve tells you purely how the market prices time, without any of the directional-fear asymmetry mixed in.

What you get for the extra work is resolution nothing else offers. Day-by-day curve points across the full chain, visible event kinks at FOMC and CPI dates (more on that below), and pure options data with no futures risk premium baked in. What you pay is tooling and maintenance: you need a platform that computes ATM IV per expiration automatically, and because ATM moves with spot, the curve has to be rebuilt regularly during active sessions. It is not a plot-once-and-walk-away chart.

The event kink

This is the payoff for building the SPX options curve: it can locate the exact calendar date where the market expects a volatility event. Because there is a curve point for every expiration, a scheduled catalyst (a Fed decision, a CPI release, a quarterly earnings report) shows up as a localized distortion in the curve.

The mechanics are simple. Options expiring before a scheduled event do not price that event at all, since it falls outside their time horizon. Options expiring after the event do. That mismatch creates a visible step-up, a "kink," precisely at the expiration that straddles the event date.

Pull up the SPX term structure the week before any major FOMC meeting and you will usually see it: a clear IV elevation at the expiration just past the announcement. That kink is a direct market estimate of how much volatility the FOMC is expected to generate, and its magnitude quantifies the event premium.

No commentary required, the curve just tells you.

The red dot marks the kink at the expiration just after the FOMC. The dashed line shows where the smooth curve would sit without the event premium. Illustrative only.

What "at-the-money" actually means when you build the curve

"At-the-money implied volatility" has two practical interpretations, and the distinction matters if you want a clean term structure.

Simple ATM: the strike closest to current spot. If SPX is at 5,720, the 5,720 call and put are your ATM options. Easiest to implement, and perfectly good for a quick read. The limitation is that the strike grid is discrete; the nearest available strike may sit 10 or 25 points from spot, which sneaks a small but real skew component into your "ATM" reading.

Delta-neutral ATM: the strike where the absolute delta of the call equals the absolute delta of the put, the point where the option is most symmetrically priced. This is the more precise definition used in professional volatility trading, and it is what most volatility surface models mean when they say "ATM vol."

For reading the overall shape of the curve (contango versus backwardation, slope steepness, event kinks) either works. The delta-neutral version is marginally cleaner and worth the effort for any quantitative work that leans on precise IV comparisons across expirations. For eyeballing the curve before the open, honestly, simple ATM is fine.

Method comparison: which to use when

Continue learning

Building the curve is the technical foundation. The companion article covers how to read its shape: why contango and backwardation transitions matter for macro positioning, and how to spot the reversal signal that actually counts. For the surrounding toolkit, the Volatility Analytics guide covers the dashboard view, Vanna & Charm covers the second-order Greeks that vol shifts drive, and 0DTE & gamma risk covers the shortest end of the curve.