Correlations closed down .89 and for the first time in 3 weeks closed under .90. The correlation surface has begun to curl on every timeframe consistent with IV dropping off.
The dynamic that we have been talking about for the past weeks, the relationship between implied volatility and historical realized volatility continued to hold this week and closed at the widest spread yet, -16.36 – with volatility selling curiously off alongside the S&P. Likewise the percentile closed at .56 and based on the amount of data we track this indicates that there have only been about 22 days with a wider spread. Unsurprisingly all of these days were in 2008, and they were pretty much clustered together for nearly 3 straight months starting in mid October 2008 and ending December 19, 2008.
So the question becomes whether or not implied volatility is “right”. Is the next month of volatility going to be significantly lower than the past month – and what does it mean for price. As our only analogue appears to be 2008, this is what that looked like:
SPY ended considerably lower by the time the spread started to normalize. Now in this case n=1 and changing market structure may mean that the relationship between implied volatility and realized volatility in 2020 and in 2008 are nothing alike. We need to keep that in mind.
In summary of this week’s post, the implied – historical volatility ratio stopped pinging back and forth violently, and started settling at continuously wide spread and low percentile. Our previous posts have mentioned that until this spread normalizes (ie. goes somewhere into the belly of the distribution) the best we can expect is extreme chop. Until we are proven otherwise (rally that sticks for more than 1-2 days) we think that remains the case this week – especially as the spread has become the widest yet of the sell off.
Hope this post was helpful and we look forward to a busy week.