The current VIX level of 26 is equal to the median VIX level over the last three recessions. As Goldman warns, while extreme VIX levels periodically occur, our analysis shows that VIX levels in the high-twenties to low-thirties for extended periods of time are rare outside of recessions. Furthermore, this was foreseeable as equities were ignoring potential warning signs from other asset classes prior to the recent sell-off.
Via Goldman Sachs,
While extreme VIX levels periodically occur, our analysis shows that VIX levels in the high- twenties to low-thirties for extended periods of time are rare outside of recessions. The second quarter revised US GDP print was 3.7% and our tracking estimate for Q3 currently stands at 2.3%, which biases us toward a mean reversion to lower VIX levels.
High-twenties to low-thirties VIX equates to recession volatility
On the economic front: Many investors have argued that VIX levels in the high-twenties to low-thirties are justified. We argue that while periodic spikes should be expected, it is hard to sustain high VIX levels outside of recessions. A few simple statistics:
- Recession volatility: The median level of S&P 500 realized volatility in a recession month has been 17.5 back to 1929.
- Non-recession volatility: The median level of realized vol in a non-recession month has been 11.4.
- VIX over the last three recessions: VIX levels go back to January 1990. Since that time there have been three recessions. Average VIX levels in the first two recessions (1990-1991, 2001) were 25 and 26 respectively. The worst of the worst was of course the Great Financial Crisis. Average VIX levels in the 2008-2009 recession were 34.
Volatility perspective: A simple VIX exercise. Suppose that we are not in a recession. Then, applying a hefty 5 point risk premium to the non-recession S&P 500 realized volatility number of 11.4, would suggest that median VIX levels might be somewhere in the neighborhood of 16.5. Even if we were in a recession, applying the same 5 point risk premium to the median level of S&P 500 realized volatility in a recession month would suggest a VIX level of 22.5 (17.5 + 5 = 22.5). Using averages instead of medians pushes the VIX higher due to the skewness in the underlying distribution of volatility (22.5+ 5=27.5).
While the VIX may have overshot its typical response to the S&P, the bigger question is whether equities were ignoring warning signs from EM FX, HY, oil and rates moves that have been building over the last year. Was the VIX too low to begin with?
Our results do suggest that relative to cross asset risk metrics the VIX was low in mid-August.
But that same analysis also shows that the VIX not only caught up but actually overshot cross asset risk moves by 12-14 points when it peaked above 40. Using cross asset risk metrics in order to generate a VIX range points to current VIX levels between 22 and 24 versus a closing VIX level of 26.05 on August 28th. While the VIX may struggle to find its new home, we think it’s headed lower if our economic projections are correct.
Were equities ignoring potential warning signs from other assets prior to the recent sell-off?
In our view, equity volatility is part cyclical, part positioning and liquidity, and partly reflective of financial distress or systemic risk. So the VIX can be high even when the economics are not poor. We tend to look at cross asset relationships during periods of stress. Cross asset risk metrics often provide higher frequency clues as to whether the VIX may be over- or under-shooting the stress in another asset classes. Earlier in 2015, non-equity asset volatility levels had been pushing higher while the VIX was still below 13 in mid-August. What if equity volatility was lagging and the VIX should have been much higher coming into the recent bout of equity volatility? That would imply that the VIX was too low and that equities were merely “catching-down”.
Which assets have historically been the most correlated to the VIX?
Emerging market weakness, the decline in oil prices, rise in high yield spreads and increase in rates volatility have all been areas of concern in 2015. Exhibit 8 shows the correlation between the VIX vs. other asset classes based upon daily changes each calendar year back to 2008. US HY spreads and EM FX have consistently been the most positively correlated to the VIX, while correlations back to S&P 500 returns have been strong and negative ranging from -0.76 to -0.89. Oil and rates have tended to have the weakest correlations back to the VIX.
Howver, Goldman concludes that in their view, the large VIX overshoot relative to recent S&P moves points to a vol spike driven more by positioning and liquidity more than fundamentals… though that merely ensures they are not antagonzing their economic ‘forecasters’ guess for growth this year. rather than fundamentals.
