“Self-similarity observed in EKGs is a potentially important gauge of the condition of our hearts.”
-Geoffrey West
If you didn’t know that “healthy hearts have relatively high fractal dimensions”, now you know. If you don’t know what a fractal dimension is, I highly encourage you to get to know about that too.
“In 1982 Mandelbrot published a highly influential and very readable semi-popular book titled The Fractal Geometry of Nature. This inspired tremendous interest in fractals by showing their ubiquity across both sciences and the natural world.” –Scale, pg 143
I’ve been trying to incorporate the fractal teachings of power-laws, non-linearity, and self-similarity to my market timing tools for many years now. There’s still so much to learn and so much to improve upon.
Back to the Global Macro Grind…
After the 6th down day (for the SP500) in the last 7, my risk management #process is registering some immediate-term TRADE #oversold signals. That’s why I have more LONGS than SHORTS in Real-Time Alerts this morning.
What’s an #oversold signal? That’s simply defined as when a market price sells off to the low-end of my @Hedgeye Risk Range. Conversely, an immediate-term #overbought signal is when the price taps the top-end of my range.
Some of our newer subscribers sometimes think of the top and low-end of my ranges as technical “resistance” and “support.” Since the range is both dynamic and always changing, that’s really not what they are.
If I change the volatility parameters in my model:
- The risk range will WIDEN with HIGHER vol inputs and
- The risk range will NARROW with LOWER vol inputs
So you do need a human with a rules-based system running the machine. This human in particular uses multiple macro factors across 3 core risk management durations (TRADE, TREND, and TAIL) to do that.
In contrast to the immediate-term and dynamic risk range, my TRADE, TREND, and TAIL levels should be thought of as critical levels of resistance and/or support.
As you can see on slide 22 of our current Macro Deck, in order to generate my TRADE (3 weeks or less), TREND (3 months or more), and TAIL (3 years or less) quantitative risk management views, I use a 3-factor model:
- PRICE
- VOLUME
- VOLATILITY
Then, in order to learn more about the most important factor in the model (VOLATILTY), I use real-time market information from the futures and options market to measure and map the “vol of vol”, or the volatility of volatility.
Back to the quantitative statement that “my risk management #process is registering some immediate-term TRADE #oversold signals” this morning, what that really means is:
- Both the SP500 (SPY) and Financials (XLF) are approaching the low-end of their respective @Hedgeye Risk Ranges
- IVOL (implied volatility) has pushed to the top-end of its recent range
One way to think about the vol of vol is where expectations of future volatility are priced.
Since The Machine delta-hedges maniacally in the 1-month price momentum window, that’s why I do a lot of measuring and mapping of those expectations looking at 1-day, 1-week, and 1-month durations.
On a 1-day snapshot, this is what implied volatility did vs. the volatility that’s been realized in the last 30-days:
- SPY’s IVOL shot up to a +31% PREMIUM vs. 30-day realized
- XLF’s IVOL ramped to a +46% PREMIUM vs. 30-day realized
Unless you’re bean-counting these moves daily, that probably sounds interesting to you, but also probably means nothing. In context, 1-month ago the IVOL on the SPY was trading at a -35% DISCOUNT to what’s been realized!
That capitulation of the SPY shorts (1 month ago) ran in conjunction with the highest net SHORT position EVER in the SP500 (looking at non-commercial CFTC futures and options contracts).
Ever, as I like to remind mean reversion people, is a very long time.
We measure and map consensus net long and short positioning across all of macro (Equities, Rates, Currencies, Commodities, etc.) weekly and contextualize it using 1-year and 3-year z-scores.
Any net long or short move that’s > 2 standard deviations on a 1-year z-score back-tests as a high probability fade (i.e. do the opposite of what the position just became).
On FEB 5, 2019, the net SHORT position in SPX (Index + Emini) was -125,139 contracts. That registered -3.96x on a 1-year z-score. Wow is all I have to say about that.
I can whine about my own ignorance in not knowing that consensus was that bearish at that point in time (due to government shut-down, we weren’t getting that data in real-time), or I can just accept it, learn from it, and move on.
In response to all of this data embedded in my decision making process, a client in Houston yesterday sat back in his chair and said “wow man, that’s a lot.”
I said, yes it is. It’s taken me many years to enrich and evolve my #process. There’s much more to be done.
Our immediate-term Global Macro Risk Ranges (with intermediate-term TREND signals in brackets) are now:
UST 10yr Yield 2.61-2.76% (bearish)
UST 2yr Yield 2.43-2.56% (bearish)
SPX 2 (neutral)
RUT 1 (bearish)
NASDAQ 7 (bullish)
Energy (XLE) 64.01-66.78 (bullish)
Shanghai Comp 2 (bullish)
VIX 13.31-17.97 (neutral)
USD 95.70-97.20 (bullish)
EUR/USD 1.12-1.14 (bearish)
Oil (WTI) 54.90-57.95 (bullish)
Gold 1 (bullish)
Best of luck out there today,
KM
Keith R. McCullough
Chief Executive Officer
