VIX 2023-2026: not a compression but a low calm twice broken by violent spikes

From November 2024 to April 2026, the VIX closed at an average of 18.9, in line with its 1990-2025 mean of 19.4. Yet the low calm in place since 2023 broke twice: on 5 August 2024 (a 38.57 close) and in April 2025 (a 52.33 close on the 8th).

“Compression” became the shorthand for this regime. Cboe’s daily closes do not support it, and the options market tells a third story: crash protection has never been this expensive on average.

A normal average, a jumpier index

Across the 384 sessions from 1 November 2024 to 30 April 2026, the VIX closed at an average of 18.9, with a median of 17.4 (Cboe daily history, Eco3min calculation). The 1990-2025 average is 19.4. The period’s low, 12.77 on 6 December 2024, sits well above the 9.14 of 3 November 2017.

The prevailing narrative since 2024 describes volatility as compressed, even crushed. The data contradict it on two counts. Only 13% of sessions closed below 15, against 32% over 1990-2025. The VIX itself moved more than usual: its 30-session rolling standard deviation averaged 2.8, against 2.1 over 1990-2025. The term structure, for its part, is unchanged: the VIX3M / VIX ratio, which sets three-month expectations against one-month ones, averaged 1.118, against 1.122 over 2009-2025.

What sets this regime apart is its breaks, not its level.

Two breaks, then a warning

The first came on 5 August 2024. The unwinding of yen-funded positions (the carry trade and its risks) sent the VIX to a 38.57 close, after a far higher intraday spike. The retreat was quick: by 16 August, the index closed at 14.8.

The second, in April 2025, lasted longer. After the 2 April 2025 US executive order imposing so-called reciprocal tariffs, the VIX closed at 46.98 on 7 April and at 52.33 on the 8th, its highest close since 1 April 2020 (57.06). Twelve closes above 30 followed one another between 3 and 22 April.

A shorter episode came in late March 2026: 31.05 on the 27th, 30.61 on the 30th. In all, the index has closed above 30 fourteen times after 5 August 2024. Over the period, the term structure inverted on 42 sessions, with the VIX above the VIX3M: the market paid more for the coming month than for the coming quarter.

What real compressions looked like

2017 is the benchmark. The VIX averaged 11.1 that year, the lowest annual level of 1990-2025, with a standard deviation of just 1.35 and a low of 9.14. The 2003-2007 phase held around 16.1 for five years, bottoming at 9.89 on 24 January 2007. The low dispersion matters as much as the low level: a market that barely moves rewards strategies that sell volatility, because their premiums come in steadily until the day they stop. Even 2019, often cited, averaged 15.4, with a low of 11.54 on 26 November.

The 2017 calm ended on 5 February 2018. The VIX went from a 17.31 close on 2 February to 37.32 on the 5th, a 2.2-fold jump. The major VIX spikes since 1990 put that jump alongside the series’ other shocks.

At an 18.9 average, the November 2024 to April 2026 period does not belong to that family.

Do same-day options escape the VIX?

The VIX measures 30-day expected volatility of the S&P 500, derived from SPX option prices, according to Cboe’s published methodology. A growing share of SPX option activity, however, sits in contracts that expire the same day. In July 2026, these 0DTE (zero days to expiration) options made up a record share, the majority of SPX options volume, according to Cboe’s volume report of 5 August 2026.

The shift continued. Cboe’s report of 5 October 2026 puts third-quarter 0DTE volume at a quarterly record, again well over half of SPX options volume (Eco3min calculation).

If part of hedging and speculation moves to same-day expiries, demand for one-month options weighs less on prices, and the VIX captures less of the stress that builds and fades within a session. A dealer who sells a same-day put can hedge it within hours, and that flow need not touch the one-month strikes from which the index is computed. August 2024 illustrates it: an intraday spike far above the 38.57 close.

The reading has a measurable limit. The volatility risk premium, the gap between the VIX and the S&P 500’s realised volatility over the next 21 sessions, averaged 4.4 points over the period, exactly its 1990-2025 average. One-month protection is not being sold cheap.

The memory of 2018 and the price of tails

The most common fear is a repeat of Volmageddon. Patrick Augustin, Ing-Haw Cheng and Ludovic Van den Bergen dissected its mechanics in “Volmageddon and the Failure of Short Volatility Products” (Financial Analysts Journal, vol. 77, no. 3, 2021). On 5 February 2018, exchange-traded products short volatility, including XIV and SVXY, lost more than 90% in a single session. To stay neutral they had to buy VIX futures, and that buying pushed futures prices higher. The authors liken the loop to the losses of portfolio insurance schemes.

Today’s market hedges against that risk, and pays for it. Cboe’s SKEW index, which prices options protecting against an extreme fall in the S&P 500, averaged 148.4 from November 2024 to April 2026, far above its 1990-2025 mean. The index rises when far out-of-the-money puts become expensive relative to at-the-money options, in other words when investors pay up for insurance against a fall of several standard deviations. No window of the same length ending before 2024 averaged as high, and the series’ daily record dates from 18 February 2025.

Volatility of volatility followed. The VVIX, which measures expected volatility of the VIX, averaged above its 2006-2025 mean. An average VIX, a high VVIX and a record SKEW describe a market at ease with the near term that pays up for the worst case.

Two caveats apply. No public data opened for this analysis sizes today’s short-volatility positions, so the 2018 analogy remains a hypothesis. And a high SKEW can also reflect scarcer supply of protection, not only stronger demand.

Since June 2026: calmer, with no inversion

The chart stops on 16 June 2026. From 17 June to 7 October 2026, the VIX closed at an average of 16.2, with a high of 20.66 on 29 July and a low of 14.21 on 22 September. It closed at 15.08 on 7 October (Cboe, Eco3min calculation). The VIX dataset carries the full daily series.

Stress gauges eased with it. The VVIX fell back below its long-run mean over those four months. The SKEW eased, while staying above its long-run mean. The VIX term structure did not invert once: the VIX3M / VIX ratio averaged 1.17, and stood at 1.175 on 7 October.

That gauge flagged every break of the period. In normal conditions, longer protection costs more because more can happen over three months than over one; an inversion means near-term fear outweighs that time value. A VIX3M / VIX ratio below 1 means the market is paying more for immediate protection than for three-month cover; it happened on 42 sessions between November 2024 and April 2026, and on none since June. For the long-run context, see the VIX as the compass of equity stress, and for credit, stress episodes compared across VIX and credit.

The regime that began in August 2024 has the average of an ordinary era and the tails of an anxious one. The open question is where the next shock starts: from a lower base, as in 2018, or from a nervous calm like this one, which has already absorbed two breaks and a warning without any prior compression.

Last updated — 8 October 2026