Learn · Options · Spreads, Volatility Structures and the Surface
VIX, VIX Futures and Volatility ETPs
The VIX is a calculation over S&P 500 option prices, so nobody can hold it; every product that offers VIX exposure holds VIX futures instead. The futures curve usually slopes upward, so a product that keeps rolling one month of futures sells cheap contracts and buys dear ones every day — in 2019 the VIX fell 46% and VXX fell about 68%. Long-volatility products pay in spikes and leak the rest of the time; inverse ones collect the leak and carry the spike.
The VIX is a calculation, not something you can own
Cboe computes the VIX continuously from the prices of a strip of S&P 500 options — out-of-the-money puts and calls across many strikes, in the two expiries that bracket 30 days — and expresses the result as an annualised volatility in percentage points. A VIX of 20 says that index option prices imply about 20% annualised volatility over the next month; divide by the square root of 12 and that is roughly 5.8% for the month itself (O12 turned implied volatility into an expected move the same way). It is the price of index protection quoted as a volatility, which is why it is called a fear gauge. Because it is a formula over option prices, there is no spot VIX to buy, store or deliver. Gold can be held in a vault and a share in an account; the VIX can only be calculated. You cannot buy it at 13 and hold it until it reaches 40. Every product that offers “VIX exposure” therefore holds something else: **VIX futures**, listed since 2004, **VIX options**, listed since 2006 and priced off the futures, or notes and funds that hold the futures for you. The index also **mean-reverts**. Since 1990 it has averaged around 19 to 20; it spikes violently and decays slowly — an intraday record of 89.53 on October 24, 2008, a record close of 82.69 on March 16, 2020 — and spends long calm stretches in the low teens. Mean reversion is why the futures do not follow the index. A future expiring in two months prices where the VIX is expected to be in two months, so when the index is low the futures sit above it, expecting it to rise back toward its average, and when the index is in a panic the futures sit below it, expecting the panic to fade. • VIX = 30-day implied volatility of the S&P 500, from a strip of SPX options. • It cannot be held: exposure means futures, options on futures, or products holding futures. • It mean-reverts: futures sit above a low VIX and below a high one. What a VIX level implies for the next month (VIX ÷ √12) — VIX 12 — a calm market: About 3.5% for the month · VIX 20 — near its long-run average: About 5.8% ← · VIX 40 — a sell-off: About 11.5% · VIX 80 — March 2020: About 23%
Futures, contango and the roll
VIX futures expire monthly and settle to a special opening quotation of the index computed from S&P 500 option prices on a Wednesday morning. Line them up by expiry and you have the VIX futures curve. Most of the time it slopes upward from the spot index through the months — **contango** — for the two reasons above: the index is usually below its average and expected to drift back up, and sellers of volatility insist on being paid for carrying crash risk, the same variance risk premium that O12 found in implied volatility. In a crisis it inverts — **backwardation** — because the spot index is high and the market expects it to fall. Now follow a product that wants a constant one month of exposure. It cannot hold one future forever, because each one expires, so it keeps selling the nearer contract and buying the next. In contango that means selling cheap and buying dear, every day. The cleanest way to see the cost is to freeze the curve. On the calm day in this lesson’s calculator the front month is 15.50 and the second month 17.00. Buy the second month at 17.00 and wait a month with nothing changing, and that contract is now the front month — priced at 15.50. That is a loss of 8.8% in a month in which the VIX did not move at all, and twelve such months compound to about −67%: almost exactly what VXX did in 2019. The roll is not always a cost. When the curve inverts, the product sells the expensive near contract and buys the cheaper one behind it, and the frozen-curve arithmetic pays: with the front at 40.00 and the second month at 34.00 the roll earns 17.6% in a month. That is the only state of the world in which a long-volatility product is paid to wait — and it is a state that does not last, because the panic that inverted the curve is exactly what the futures expect to fade. • Contango (front below back): the normal state — the roll costs the holder. • Backwardation (front above back): the crisis state — the roll pays. • Monthly roll cost with a frozen curve ≈ front price ÷ second-month price − 1. The roll on two days (this lesson’s calculator) — Calm: spot 13.00, front 15.50, second 17.00: Contango of 9.7% between the months · One month, curve unchanged: −8.8%, with the VIX flat · Twelve such months: About −67% ← · Crisis: front 40.00, second 34.00: +17.6% in a month
The products, and why the long ones decay
The original **VXX** — an exchange-traded note tracking the S&P 500 VIX Short-Term Futures Index, which holds the first two monthly futures and rolls between them to keep about one month of maturity — launched in January 2009 and matured in January 2019, having lost more than 99% of its value; its successor has traded under the same ticker since. On a long chart the losses are disguised by repeated reverse splits, which is why the replay in this lesson shows split-adjusted prices in the thousands of dollars. The decay is not a fee or a flaw in the product; it is the roll in contango, paid faithfully, for years. Two other families are built on the same futures. **Leveraged** funds such as UVXY aim for a multiple of the futures index’s daily return — 1.5×, cut from 2× after February 2018 — and **inverse** funds such as SVXY aim for the opposite of it — −0.5×, cut from −1× at the same time. The word that matters in both is *daily*. A product that resets its leverage every day does not deliver the multiple of the index’s return over a month; it compounds the multiple of each day’s return, and in a volatile market that drags. If the futures index rises 10% and then falls 10%, it ends down 1%; a 2× daily product ends 1.2 × 0.8 = 0.96, down 4%. Inverse products are the mirror image, and that is their appeal: they collect the roll that long products pay, which is why they rise steadily through calm years. What they sell in return is the spike. A −1× product loses everything if the futures index doubles in one day, and on February 5, 2018 the largest inverse note came close enough to that to be terminated — the case O21 reconstructs. The leverage cuts that followed were the industry’s admission that the tail was larger than the products had been sized for. • Long (VXX): pays the roll in contango — decays through calm years. • Leveraged long (UVXY, 1.5× daily): pays the roll faster and suffers volatility drag. • Inverse (SVXY, −0.5× daily): collects the roll and is short the spike. • Daily reset: the multiple applies to each day, not to the month. Daily reset over two days (futures index +10%, then −10%) — The futures index: 1.10 × 0.90 = 0.99 — down 1% · A 2× daily product: 1.20 × 0.80 = 0.96 — down 4% ← · A −1× daily product: 0.90 × 1.10 = 0.99 — down 1%, though the index also fell
The year long volatility paid: 2020
If the roll is the cost, the spike is the payoff, and 2020 delivered the largest one on record. On February 14 the VIX closed at 13.68 and VXX at a split-adjusted 865. By March 16 the VIX had closed at 82.69, its highest close ever, up about 500%; VXX made its highest close two sessions later, at 4,416, up about 410%. The note rose a great deal — and still less than the index, because its futures priced where the VIX was expected to be a month out, after the panic had partly faded. For a few weeks the curve sat in backwardation and the roll paid. Then it went back to costing. By the end of the year the VIX was still 66% above its mid-February level, at 22.75, while VXX was only 24% above, at 1,075. From June 30 to December 31 alone the VIX drifted down 25% and VXX fell 51% — the roll in contango, back at work on a curve that stayed steep because the market remained nervous. From the last close of 2019 to the last close of 2020, the most turbulent year since 2008, the VIX rose about two-thirds and VXX about a tenth. That is the timing problem in three parts. Owned before the spike, a long-volatility product costs the roll every month, often for years. Bought during the spike, it is bought after most of the move. Held after the spike, it gives the gain back as the curve returns to contango. None of that makes it useless — a short holding through a known risk window can be a sensible hedge — but it does make it a trade rather than an investment, and a trade that has to be sized in advance for its monthly cost. 2020 on the daily series (VXX split-adjusted) — February 14: VIX 13.68 · VXX 865 · Peak closes: VIX March 16, VXX March 18: VIX 82.69 (+504%) · VXX 4,416 (+410%) · June 30 → December 31: VIX −25% · VXX −51% ← · Full year, from the last close of 2019: VIX +65% · VXX +11% Every figure on this page comes from the same daily closes the replay below fetches; the replay asks you to call two of them before you see them.
Using volatility products without being used by them
As a hedge, a long-volatility product has to be judged the way O7 judged a protective put: what does the protection cost per period, and can the position afford it every period it is renewed? In contango that cost is the roll — several percent a month on a calm curve — so the defensible uses are short and specific: through a known risk window, sized so that the monthly leak is an affordable premium. For most portfolios an index put has the cleaner economics, because its cost is fixed when it is bought and it does not roll. Inverse products need the opposite discipline. They are paid the roll, which feels like income, and they are short exactly the spike that long products exist for. Size them as if the spike happens tomorrow, because the day it happens is not announced and the daily reset means there is no waiting for a recovery — a product can be terminated before the futures fall back. And for every product, read the prospectus before the price: which index it tracks, the leverage and the daily reset, and the early-redemption or acceleration terms that let the issuer close it after a large loss. The summary rule is simple to state and easy to forget. A VIX product is not an investment in fear. It is a position on the VIX futures curve — on its level, its slope and how both move — and its return is decided as much by the slope as by the VIX itself. • Long volatility as a hedge: short holding periods, sized for the monthly roll. • Inverse volatility: size for tomorrow’s spike; there is no waiting out a daily-reset loss. • Read the prospectus: index, leverage, reset, and termination terms. • Every VIX product is a position on the futures curve, not on the index. A steady rise in an inverse volatility product is the roll being paid to you for carrying the spike. The steadier the rise, the larger the position tends to grow — which is the failure mode O21 describes.
What you'll practise
Why can’t a fund simply hold the VIX?
40 XP in the app · multi select
Sources
- How the VIX is calculatedCboe Global Markets, VIX White Paper
- VIX futures contract specifications and settlementCboe Futures Exchange, VIX futures specifications
- The index behind VXXS&P Dow Jones Indices, S&P 500 VIX Short-Term Futures Index methodology
- Leverage changes to volatility funds after February 2018ProShares, UVXY and SVXY prospectus supplements (2018)
- VIX record closesCboe VIX historical data
Learn content is for education only — not individualized financial advice, a recommendation, or a solicitation to buy or sell any security. Options involve substantial risk. Examples are simplified and historical patterns never guarantee future results.