Traders read a jump in the 9-day VIX, while the 30-day stays calm, as the first tremor before broader volatility. Across 3,876 days, when VIX9D outran VIX the 30-day gauge fell, not rose.
The VIX everyone quotes measures expected volatility over the next 30 days. Its lesser-known sibling, VIX9D, measures the next nine. Because the shorter window reacts faster, a piece of folklore has grown up around it: when VIX9D jumps while the 30-day VIX stays calm, near-term fear is the first tremor, and broader volatility is about to follow. The 9-day is the canary in the coal mine. We tested that on every trading day since VIX9D's record begins, 3,876 days from 2011 to today. The canary mostly cries wolf.
The first problem with the story is that the two gauges almost never diverge. Day to day, VIX9D and the 30-day VIX move together with a correlation of 0.93. In fifteen years, VIX9D jumped 10% or more while the 30-day stayed nearly flat (up less than 3%) on just 59 days, about four times a year. The clean setup, the canary chirping while the mine is quiet, is rare to begin with, because near-term and 30-day fear are mostly the same fear measured over two windows.
To get a real sample, we flagged every day VIX9D's move outpaced the 30-day VIX's by a top-decile margin, 367 days in all. If the canary worked, the 30-day VIX should climb over the following weeks. It fell.
| Days after the canary | 30-day VIX change | VIX higher | Baseline (any day) |
|---|---|---|---|
| 5 days | -0.33 pts | 41% | -0.17 pts / 46% |
| 10 days | -0.58 pts | 41% | -0.19 pts / 46% |
| 20 days | -0.56 pts | 43% | -0.28 pts / 46% |
In the ten days after the canary, the 30-day VIX fell a median 0.6 points and was higher just 41% of the time, against 46% on a random day. The early warning was followed by slightly less broad volatility than average, not more.
Push the divergence to its extreme, the 921 days the short end actually inverted (VIX9D closed above the 30-day VIX), and the result only sharpens.
| Days after the inversion | 30-day VIX change | VIX higher |
|---|---|---|
| 5 days | -0.98 pts | 36% |
| 10 days | -1.96 pts | 31% |
| 20 days | -2.38 pts | 30% |
After an inversion the 30-day VIX fell a median 2 points over ten days and was higher only 31% of the time. And note where these days sit: the average VIX on an inversion day was 21.9, not a calm tape. The short end inverts during stress that is already underway, not before a quiet market cracks.
Here is the mechanism that kills the thesis. After the canary, the gap between the two gauges does close, but by the near-term number coming down, not the broad one going up.
| In the 10 days after the canary | |
|---|---|
| VIX9D (the 9-day) change | -1.1 pts (falls) |
| 30-day VIX change | -0.6 pts (falls less) |
| Broad-vol spike rate after a canary | 20% |
| Broad-vol spike rate, any day at the same VIX | 18% |
The short end simply relaxes back toward the long end. The fear the canary registered passes, rather than spreading. And what little follow-through exists is not the divergence at all, it is the VIX level: a broad-volatility spike (a 5-point VIX rise within ten days) showed up after 20% of canaries, against 17% on a random day, a gap that all but vanishes once you compare like for like, since any day at the same VIX level produced 18%. The canary tells you almost nothing a glance at the VIX itself does not. Like the SKEW index, it is a volatility gauge that looks predictive and is not.
A faster gauge reacts sooner, but reacting is not predicting. When VIX9D outran the 30-day VIX, broader volatility followed no more often than on any ordinary day, and the gap closed by the canary falling silent.
Every trading day since VIX9D's record begins, 3,876 days from 2011, with the 9-day and 30-day VIX tracked against the S&P 500.
VIX9D and the 30-day VIX move together with a 0.93 correlation. In 15 years VIX9D jumped alone, while VIX stayed flat, just 59 times. They rarely diverge at all.
In the 10 days after VIX9D outran VIX, the 30-day gauge fell a median 0.6 points and was higher just 41% of the time, against 46% on a random day.
After the signal, VIX9D dropped 1.1 points while the 30-day VIX dropped 0.6. The divergence resolved by near-term fear fading, not broad volatility arriving.
A broad-vol spike followed 20% of canaries, against 18% for any day at the same VIX. The divergence added nothing a look at the VIX did not already give.
Why the faster volatility gauge is not the wiser one.
VIX9D reacts sooner because its window is shorter, not because it knows more. Reacting first and predicting are different things, and folklore tends to confuse the two.
VIX9D and VIX move together 93% of the time. A signal built on them diverging is rare by construction, and rare signals are easy to remember and hard to actually trade.
The canary's only follow-through was the VIX level hiding inside it. Compare any signal to a plain day in the same conditions before crediting it with foresight.
We test popular signals the honest way: every instance counted, every result measured against a plain baseline. See what else held up, and what did not.
Browse Reality Check
Comments
Loading comments…