Update: read sequel article here.
There has been some recent murmurs about what volatility implies for the stock market. This happens, without coincidence, as this month we see both markets streak to record highs, and volatility embracing multi-year lows. What we argue below is that while there is some general inverse correlation between markets and volatility (i.e., rising markets equates to falling volatility), the relationship at the extremes is far stronger than it is for the broad aggregate. And it’s true (previous article here) that -smoothly across time- at these low levels of volatility, both the absolute return and the risk-reward ratio for the market is undoubtedly unfavorable.
Now let's take a closer look at the subsequent 2 weeks (10 trading days) performance chart of the S&P based on vigintiles of the VIX. The median levels part-way through this time is at 6 days, after each VIX ventile noted in red text. The noise about these levels is reasonably high (and we will soon demonstrate this). Notice in the upper left of the chart below (vigintiles 1&2 versus say 19&20) that markets with VIX falling below 12% are different from markets with VIX rising above 29%?
There has been some recent murmurs about what volatility implies for the stock market. This happens, without coincidence, as this month we see both markets streak to record highs, and volatility embracing multi-year lows. What we argue below is that while there is some general inverse correlation between markets and volatility (i.e., rising markets equates to falling volatility), the relationship at the extremes is far stronger than it is for the broad aggregate. And it’s true (previous article here) that -smoothly across time- at these low levels of volatility, both the absolute return and the risk-reward ratio for the market is undoubtedly unfavorable.
First see this article, the most popular for several days.
In it we are cited using the same academic research shown in the first link above. Here we argue -combined with research here and here- that record low volatility (similar to the 11's level we are at now that is generally bottom 5% or bottom vigintile of VIX) represents a future period of slow and relatively erratic market behavior, which can also include extreme downward risk in the coming months. Alternatively it is unlikely to result in a sudden shift higher
in the markets.
Second notice that the volatility has been at a multi-year low recently, closing below 11.5% a few times this month. This has not occurred in 2 years, and rarely will it ever (per black shaded data markers in the chart to the right) close slightly lower (below 10.5% in any given month).
Third we notice a strategic headline article this month as well, on the cover of Wall Street Journal's MarketWatch.
See a picture of it below, as it would mislead you to deduce that there is simply no difference
between the market returns regardless of the VIX level.
While powerful and worthy of a read, there are some
new ways one can consider the topics and analogies in Hulbert’s astute article. Let's note some of them here:
- Using the S&P which tracks better to the VIX versus the Wilshire 5000
- Changing the analysis to incorporate a variable term horizon instead of a fixed time horizon, since we know it varies partially based upon the level of the VIX itself
- Consider not only the average return, but the the dispersion of risk (or the confidence bars) about that average results
- Meditate on how changing volatility regimes impacts the results
- Not just look at select groupings of quantiles, but rather show all of them mutually exclusively
Now let's take a closer look at the subsequent 2 weeks (10 trading days) performance chart of the S&P based on vigintiles of the VIX. The median levels part-way through this time is at 6 days, after each VIX ventile noted in red text. The noise about these levels is reasonably high (and we will soon demonstrate this). Notice in the upper left of the chart below (vigintiles 1&2 versus say 19&20) that markets with VIX falling below 12% are different from markets with VIX rising above 29%?
Now let’s contrast all of these red ventile performance levels with the corresponding volatility of the returns at that point. We re-label just these red vigintiles into the chart below. So that yellow is now for the extreme low VIX we are currently in, and blue is for extreme high VIX such as during the China’s Black Monday last year and the 2008-era financial crisis (see here). It should be far more easy to grasp that at these low VIX that we have today, there is simply a ghastly return (y-axis) per unit of market volatility (x-axis).
Last we can see how sensitive our results above are, relative to the volatility regime we are in. In other words, does the central bank monetary easing in recent years invalidate the relationships we show above? As we can see in pink colored bars below, the results are not impacted very much by time relative to the other risk considerations we show above.
The conclusion with this research is that when we see statistical ideas presented, we should always vigilantly exercise caution. Don't continuously rely on any one measure or "simply trick" for an impulsive investment decision. Your financial security hangs in the balance with your choices, and so one should deliberate multiple vantage points. Not just analyze something such as VIX, in sequestration. Tread with care as there is colossal risks likely in the autumnal months ahead.
August 21 update: On Wednesday August 17 we create 3 exclusive vigintile (5% sections) for the bottom 15% of VIX (e.g., ranks 0-5%, 5-%-10%, 10%-15%). See below, and now the boxed data on the left for 8/2016 (this month) is a count of 5 so far, up from 3 shown on August 17.
See 2016 on the vertical axis, and month 8 on the horizontal axis. We see above that it's been a couple years since we had a month where the VIX closed below 11.6% as where we have it today! Just to be clear, the middle sectional panel is a count of all days in the month where the VIX closed was between 11.6% to 12.2%, and the right panel is where VIX closed between 12.2% to 128%.
What continues to be obvious from these panels is that the regime of steadily rising markets is slowly calming down, late July historically is the trough for these low-volatility occurrences that we now see, and expect remarkable market risks to ensue by year-end.
The conclusion with this research is that when we see statistical ideas presented, we should always vigilantly exercise caution. Don't continuously rely on any one measure or "simply trick" for an impulsive investment decision. Your financial security hangs in the balance with your choices, and so one should deliberate multiple vantage points. Not just analyze something such as VIX, in sequestration. Tread with care as there is colossal risks likely in the autumnal months ahead.
August 21 update: On Wednesday August 17 we create 3 exclusive vigintile (5% sections) for the bottom 15% of VIX (e.g., ranks 0-5%, 5-%-10%, 10%-15%). See below, and now the boxed data on the left for 8/2016 (this month) is a count of 5 so far, up from 3 shown on August 17.
See 2016 on the vertical axis, and month 8 on the horizontal axis. We see above that it's been a couple years since we had a month where the VIX closed below 11.6% as where we have it today! Just to be clear, the middle sectional panel is a count of all days in the month where the VIX closed was between 11.6% to 12.2%, and the right panel is where VIX closed between 12.2% to 128%.
What continues to be obvious from these panels is that the regime of steadily rising markets is slowly calming down, late July historically is the trough for these low-volatility occurrences that we now see, and expect remarkable market risks to ensue by year-end.
No comments:
Post a Comment