In recent months, BofA notes that the speed of mean reversion in the VIX has been particularly striking by historical standards. Since the end of QE3, VIX spikes have had very little persistence, generating low cumulative volatility relative to the previous 25 years, underscoring BofA's thesis of a fragile market that features rapid jumps from states of calm to states of stress and back.
As BofA details, markets are hyper-sensitive today to central bank action / rhetoric, with the beta of global equity, commodity, fixed income, FX, and corporate credit markets to 10yr US Treasuries near 26+ year highs.
Hence it is perhaps not surprising that as 10yr Treasury yields fell swiftly following last week’s FOMC decision to not raise rates, the S&P 500 rallied back to 2180 (~35bps away from its pre-selloff high) and the VIX dropped under 12 (to within 0.1 vol points of its pre-spike low).
However, the speed of mean reversion in the VIX has been particularly striking by historical standards. Chart 7 (above) plots the relationship between the magnitude of a VIX spike and the total amount of volatility subsequently generated before the VIX retraces back to near pre-spike levels. It shows that since Oct-14, VIX spikes have had very little persistence, generating low cumulative volatility relative to how comparably sized spikes evolved from Jan-90 to Oct-14. Moreover, the recent shock was particularly fleeting, recording the smallest total amount of volatility (i.e., lowest y-axis reading).
At the same time, VIX spikes above 20 have occurred twice as frequently on average over the past two years (Chart 8 above) and 2016 is on track to see the largest number of “vol events” (rise of > 25%) since 1990 for the VIX...
Underscoring our thesis of a fragile market that features rapid jumps from states of calm to states of stress and back.
* * *
With the VIX again depressed vs. cross-asset risk & correlation ahead of a seasonally volatile period featuring US presidential elections, BofA thinks it’s prudent to reload on smart hedges:
1. Our analysis of S&P 500 implied volatility in election and non-election years since 1990 (Chart 9 below) shows support for (and even a slight increase in) both Oct and Nov expiry SPX variance swaps in election years compared to declines in non-election years. This is a stronger result than we previously established using spot VIX, as it accounts for historical term structure rolldown, and lends further support to the idea of selling short-dated puts on the VIX to help cheapen portfolio protection.
2. S&P 500 put skew remains structurally steep, with the cost of 3M 25-delta puts atthe highest on record (data since Jan-04) vs. the cost of 3M 50-delta puts (Chart 10). We like SPX Dec put spreads to capture potential catalysts beyond the US presidential election (e.g., Italian referendum or contagion from European Banks).
3. For investors seeking more downside convexity than SPX put spreads provide, we continue to recommend ratioed put calendars on the S&P, i.e. selling one longerdated OTM put to fund multiple shorter-dated closer-to-ATM puts – a trade that leverages steep term structure and is well-suited to hedging “fragility events” that tend to feature violent flattening or inversion of term structure. For example, the ratio of SPX 6M 20-delta puts to 1M 40-delta puts is currently in the 99th percentile over the past 15 years of daily data (Chart 11).