Thread by @couletsis

1/n S&P500 1month ATM volatility at 22.3 implies ~1.4% daily move for the $SPX in the next month. That’s what statistics tells us for a sole time-series that is an index. But there’s more going on under the hood.

2/n ATM option vol mirrors investor expectations for volatility realized around the current level of the market. If you want a concrete estimate for future index volatility regardless of price levels, then it’s actually variance swaps you need. Enter, the $VIX index.

3/n Irrespective of its mainstream adopted alias as a volatility index, the $VIX is -for every practical purpose- the price of a 30day variance swap on the #S&P500. And it trades somewhat above the ATM volatility, to offset for skew and convexity.

4/n In the presence of skew and convexity, the IV of the options included in the VIX calculation is higher than the ATM IV, thus the VIX prints higher than the current SPX ATM vol. It’s the price you pay for losing less and making more money if volatility drops or increases.

5/n $VIX at 28 implies ~1.7% move for the $SPX. Take note though, index volatility is actually: Weighted Average Volatility of Its Constituents multiplied by the square root of the index components correlation.

6/n In a tech-heavy index, 1month ATM vol for all its top-8 firms, that make-up for ¼ of the index cap, trade >30% and a still not-so-elevated correlation, we can see how 2.8% daily moves and a correlation of 0.4 does not still reflect some kind of alarm.

7/n Even for all $SPX components’ volatility remaining unchanged, if correlations spike from 0.4 to 0.6 then $VIX would print 34%. (If they drop back to 0.3, then VIX would dip to 22%).

8/8 You can add salt to the wound and skip a Gaussian distribution and consider power-law or heavy tail Levy alternatives and reach even lower numbers of future volatility estimates. Although many fragility gauges point to considerable nervousness, this is still not a vol event.