Taming Wild Components

How the new dispersion index DSPX fits into trading strategies

Created with Midjourney

Setting the Table:

• Once again inflation data rules the tape as slightly hotter than expected price index data sent the market lower on the day after a cobbling together string of up days. Markets have opened lightly green.

• Options markets were more active than earlier in the week with almost 47M contracts trading hands, fueled by strong SPX volume. It’s interesting to see how the LIQ index has settled into a fairly tight range for SPY.

  

  • The implied volatility measurement of market width has been much more stable than it was during the summer. This suggests that despite somewhat lower volumes and open interest, markets are maintaining their depth.

    

Identify:

• Just two weeks ago CBOE in conjunction with S&P Dow Jones launched a new index called DSPX, which tracks the relationship between individual component stocks and the expected variance in the index. 

  • Conceptually dispersion is a risk measure - how much more or less are individual components going to move then their average. Their average is of course the index itself. 

  • The base expectation for this index having a positive value is the justification for diversification. We expect that the offsetting movements of many stocks will net in a beneficial direction for both returns (up) and volatility (down). There will be a “premium” to the sum of the components because of this. 

• Comparing two different periods, we can see how the cross sectional distributions differ significantly. In both of these periods in 2021, we saw single digit percentage rises in the S&P 500, but February saw significantly different variations across components.

  

• The objective of the DSPX index is to use component VIX calculations to track what the market’s forward looking expectations of return distributions (dispersion) are. What’s important about the VIX calculation is that it uses a full strip of options to calculate distributions - not just ATM - so it incorporates measures of skewness and kurtosis. 

  • The calculation nets the difference between weighted component variance and the SPX “original” VIX, floored at 0. Components necessarily will trade higher, with the only unique case being February 2018 when VIX ETPs created a dynamic that spiked VIX futures without spiking component vol. 

• Dispersion is particularly interesting for investors to track, because it digs into the detail of “how” the market is performing. Are we seeing bottom up breadth across all sectors, or top line breakouts driving index performance?