What Exactly Does the VIX Tell Us?

Most investors know of the VIX Index, but not as many understand what information the VIX provides an investor. Here is my attempt to provide an initial outline of what it is and why the information embedded within the figure is so powerful.

VIX Defined

The CBOE has a white paper that provides a ton of detail into the VIX calculation, but the origin of the index describes what it is at a high level well enough:

In 1993, the Chicago Board Options Exchange® (CBOE®) introduced the CBOE Volatility Index® (VIX® Index), which was originally designed to measure the market’s expectation of 30-day volatility implied by at-the-money S&P 100® Index (OEX® Index) option prices. The VIX Index soon became the premier benchmark for U.S. stock market volatility.

Ten years later in 2003, CBOE together with Goldman Sachs, updated the VIX to reflect a new way to measure expected volatility, one that continues to be widely used by financial theorists, risk managers and volatility traders alike. The new VIX is based on the S&P 500® Index (SPXSM), the core index for U.S. equities, and estimates expected volatility by averaging the weighted prices of SPX puts and calls over a wide range of strike prices.

In summary... the VIX is a reflection of the market's expectation of future market volatility.


The Math Behind What the VIX Level Means

As Eddy Elfenbein of the great Crossing Wall Street blog outlined back in 2012, you can take the current market's expectation of future market volatility (i.e. the VIX), to determine the expected range of future stock market returns.

Fun fact: Take whatever today's $VIX is. Divide it by 3.46. That's the market's view of the 1 stand dev range +/- for the next 30 days. $$ 

— Eddy Elfenbein (@EddyElfenbein) March 7, 2012

The 3.46 denominator is simply the square root of 12, which takes the VIX from an annualized figure to a monthly figure given there are 12 months in a year.

You can also vary the blue "confidence bands" in the chart above (in the chart above, the one standard deviation bands "should" capture 68% of outcomes). For example, from statistics we know that 1.96 is the z value of a 95% confidence interval. We can convert this 1.96 level to a monthly value as follows:

1.96 / (square root of 12) = 0.57

We can then take that 0.57 value and multiply it by the VIX to determine the bands that (in theory) contain that 95% percent of outcomes. For example, with a VIX of 20, the equation is:

20 x 0.57 = 11.3 or 95% of all outcomes over the next month should be within a +/- 11.3% return

In reality, the VIX typically overstates the level of market risk and understates the results within the band. As the chart shows below, the confidence bands described by the VIX understate the "capture rate" of the bands, especially at higher confidence levels. In the chart above, one standard deviation bands should capture 68% of outcomes, but instead have captured 85% of outcomes.

Conclusions

Neither of the following conclusions should surprise readers of the blog, but they are:

• Takeaway 1:the VIX has done a great job of predicting the future range of market moves (outlined previously in my post The Case for a Steady Volatility-State Managed Portfolio)
• Takeaway 2: the VIX does typically overstate the future level of volatility (outlined previously in my post The Case for Put Writing / Further Improving PutWrite Performance)

As a result, while an investor can utilize the VIX to scale equity weights to smooth out returns, an investor should rarely be buying volatility protection.

Utilizing the Value of Value to Make Value / Growth Tilts

Back in August I outlined why I thought the plain-vanilla value premium had been compressed to the point growth had and was likely to continue to outperform in my post Death of (Plain Vanilla) Value - Long Live GARP. This post is meant as a follow up and suggests a few frameworks as to how an investor might allocate based on the given "value of value".


Backdrop: Value of Value Matters

Value historically outperforms growth when stocks making up the value indices are beaten down relative to growth. Thus, it should be no surprise that value materially outperformed following the Internet bubble as value stocks were massively cheap relative to growth stocks (see below). The issue over the last decade plus is that many investors have piled into value ignoring the driver of value's historical outperformance, resulting in a "value discount" that is historically narrow.

The next chart outlines in more detail why the discount matters, with the starting "value discount" on the x-axis and the subsequent 7-year excess return on the y-axis. You can see the linear relationship between the "value of value" and future value excess return to growth. When the discount is high, outperformance of value vs growth is likely. When the discount is low (which happens to be where we currently sit), underperformance of value vs. growth is likely.

Putting the Insight into Action

While there are a multitude of ways this insight can be put to use, below are a few simplistic ways that only require making a reallocation at most quarterly and has been more likely to require a reallocation once every 3-5 years. Note this analysis does have a huge data mining issue as we know in advance that a 25% and 30% value discount are thresholds where growth and value have diverged in the past - though less of an issue today as we are near all-time tight levels. As an aside, as the charts below show it wasn't until the late 1990's that the performance of the Russell 3000 Value and Growth indices diverged. It was only when investors initially flocked to growth (and later value) that we have seen distinct differences in the "value discount" and in subsequent performance.


Model 1: Long-Only

Rules (reallocate the portfolio quarterly - ignores transaction costs):

• If the value discount is narrower than -25% (i.e. growth is cheap), allocate to the Russell 3000 Growth Index
• If the value discount is between -30% and -25%, allocate to the Russell 3000 Index (i.e. don't tilt growth or value)
• If the value discount is wider than -30% (i.e. growth is expensive), allocate to the Russell 3000 Value Index

Model 2: Long-Short

Rules (reallocate the portfolio quarterly - ignores transaction costs):

• If the value discount is narrower than -25% (i.e. growth is cheap), long position in the Russell 3000 Growth Index / short position in the Russell 3000 Value index overlayed on the Russell 3000 Index (i.e. generate alpha on top of the Russell 3000 Index through the long/short)
• If the value discount is between -30% and -25%, allocate to the Russell 3000 Index (i.e. don't tilt growth or value)
• If the value discount is wider than -30% (i.e. growth is expensive), long position in the Russell 3000 Value Index / short position in the Russell 3000 Growth index overlayed on the Russell 3000 Index

The Relationship Between High Yields and High Yield / Stock Performance

I've previously posted my broader thoughts on high yield (that there is typically limited to no benefit vs. a stock / bond allocation), but the below chart provides some additional context I thought worth sharing.


Starting Yield: Anchor and Cap for High Yield Returns

The left hand chart breaks out five year forward returns vs. various starting yield-to-worst "YTW" buckets of the Barclays High Yield index (along with what the average starting yield was for the index within that bucket over the time frame). It should be no surprise that the average starting yield anchors what the return will be (when yields are low, returns are low), while the right hand chart outlines the average underperformance vs. the starting yield has been roughly 2.0-2.5% / year, irrespective of the starting yield (high yield isn't called junk for nothing).


Stocks vs High Yield Performance at Various Yields

Both charts show that equity market performance is highly correlated with the credit premium; when the YTW of high yield bonds is high, it is highly likely that the equity premium is high too - leading to higher equity returns. On the other hand, when yields are low and the credit premium (as well as equity premium) is low, stock and high yield returns are more muted. However, the unlimited potential of stocks vs the capped upside of high yield at low rates has lead to consistent equity outperformance when yields are low, while high yield has performed exceptionally after yields were completely blown out (post the 2008/09 financial crisis).

At the current yield to worst of 7.6% and a spread of almost 6% to treasuries, we should expect returns to be no more than 4.5-5.5% over the next five years. Nothing special for the potential risk, but much better than the sub 5% yield we saw a bit more than a year ago.

Using "Normal" Drawdowns as a Timing Signal

The below analysis was purely an accident. I was actually looking into periods the U.S. stock market "suffered" a 10% drawdown for the absolute opposite reason; to show that a buy and hold investor should likely ignore these regularly occurring events. How regular?

The always interesting Ryan Detrick points out:

I looked at every calendar year since 1960 and looked at various correction levels. Turns out 94% of all calendar years see at least a 5% correction, while 53% of all years see a 10% correction. Maybe this recent 12% correction isn’t so alarming?

So... what if an investor were to sell-out of their U.S. stock allocation and shift it into bonds whenever one of these standard 10% drawdowns occurred and were only willing to go back into stocks when they once again were within 10% of its all-time high?

Rules:

• At month-end if the drawdown from its previous peak was less than -10%, bonds (US Agg)
• Otherwise, stocks (S&P 500)

Well going back to 1976 (the inception of the Barclays U.S. Agg bond index), the results were surprising. Almost 100% of the return, about 20% less volatility, and less than half of the drawdown. As interesting is how well it kept up with the stock market in all periods except those that followed severe drawdowns. This is largely due to the rarity of which 10% drawdowns become 30, 40, or even 50% drawdowns, so the strategy was at most times fully invested.