Where Are We In The Volatility Cycle?

First, to understand the idea of cyclical volatility, we only need to look at a long-term chart of volatility. VIX will do. We can characterize a complete volatility cycle in four stages: the large initial shock, smaller aftershocks, a steep decline, and a period of stability.

fig. 1. CBOE Volatility Index, 2005-2012. Source: CBOE

I don't think it is helpful to parse this idea of a volatility cycle in too much detail, because ultimately this is just a heuristic that may easily be proven inaccurate in the future. We might see a series of small or moderate volatility spikes punctuating a long period, for instance, or any number of other scenarios. But with that caveat out of the way, we can ask whether markets are still working off the effects of the 2008-2010 period, or whether we are fully stable and ready for the next upturn in volatility.

fig. 2. Four Stages of Stabilization in Equity Vol Market. Source: Goldman Sachs

With short-term implied volatility (IV) well below its historical average, it might seem obvious that markets have fully normalized. However, in a 2011 report, strategists at Goldman Sachs explained why looking at short term measures like VIX is not enough. They identified four stages of equity market stabilization:

1. Realized volatility trends lower as the economic outlook improves;
2. Shorter-dated implieds track realized volatility lower;
3. The term structure becomes steeply upward sloping;
4. Longer-dated implied volatility levels decline.

Stages one and two have been complete for some time. Three month realized volatility for the S&P 500 is now below 12% and has been followed lower by option prices. Fig. 3. gives a cleaner picture of the volatility shocks we have seen in recent years and of how quickly markets have fallen back toward average levels.

fig. 3. S&P 500 3M Yang-Zhang Volatility. Source: Condor Options

To confirm the third stage, we should look not just at the slope of the current IV term structure but also how it relates to historical minimum and maximum levels. Fig. 4 includes several plots showing the maximum, minimum and median levels of IV at various time horizons (i.e. how long each option estimate is from expiration) from 2005-2012. We put all those curves together in a volatility cone to get a sense of how widely the SPX term structure can range.

fig. 4. SPX IV Quantiles by Time Horizon, 2005-2012. Source: Condor Options

fig. 5. SPX Implied Volatility Cone, 2005-2012. Source: Condor Options

Finally, we're in a position to compare current SPX IV levels to those historical ranges. The current level of 30-day SPX IV is about halfway between its absolute minimum and long-term median levels, while at a one year horizon, IV is not far from the median value. The comparison to the median tells us that the current term structure is steeper than usual, which is something we wouldn't be able to confirm from looking at current values in isolation. (Also, we're intentionally not looking at a plot of the VIX futures term structure to avoid any noise or other volatility inputs that might be reflected there.) So it looks like the third stabilization condition has been met.

The final condition is more difficult to evaluate, but also seems to be confirmed. Evidence that long-term IV has declined sufficiently can be had by comparing the two-year part of the current curve in fig. 5 to the median: options are priced just slightly below the long-run averages, and we are obviously well below the 75% quantile values. A year ago, two year SPX variance swaps were being struck near 24%, so current estimates near 21% are a sign or continued improvement. At the same time, we can't help but think that long-dated IV is unreasonably sticky.

What this means for investors is that the risk scenarios that pop in and out of headlines are not, no matter how they may seem, as significant economically as they were even a year or two ago. Tail risks will always be with us, but we have clearly completed the process of equity market volatility stabilization.