web analytics

I ran a quick study this morning to look at what happens to stock market volatility following large single day volatile events. Note that we are looking at market volatility here, not direction, because the directional message is simpler and clear: historically, large volatile events in stocks have come after declines. Since stocks mean revert, these have quite consistently been good buying opportunities. That’s a surprising message for the people who would have you think the VIX, the FEAR GAUGE, is the harbinger of doom–when the VIX goes up a lot you should be buying stocks.

I defined volatility events in three ways: as large single day declines in the S&P 500 cash index (using SigmaSpikes as a measure), as large one day point increases in the VIX or 20 day historical volatility. Charts attached to this post show a few interesting points:

  • Shocks in volatility tend to decay. One analogy is the ripples in a pond after you throw a large rock in the middle–volatility behaves something like this. (And this is an element caught in GARCH models.)
  • How we measure those volatility events does not seem to matter a lot.
  • There does appear to be a short-term increase in historical volatility, at the same time the VIX decays over the near term. Options traders might find something useful there.

A few counterpoints and things to be aware of:

  • Sample sizes are small, though the effects do appear to be persistent over time.
  • These charts just show averages and hide a wide range of outcomes. If you are doing the same work yourself you might also segregate charts into events which happen in low and high vol regimes.
  • All measures of volatility are subject to various measuring artifacts. It’s entirely possible the “knuckle” in HVol is mostly due to the right edge of the 20 day evaluation window.

Just some food for thought that might give us some guidelines to the weeks ahead.

 

spvol-3s-2010

spvol-3hvol-2010

spvol-vix-spike