Pursuing Data

“If it’s in the data, pursue it.”

-Cliff Asness

That was a great quote from a great interview that Lasse Heje Pedersen did with AQR Capital founder (and former head of quant research @Goldman) Cliff Asness in Efficiently Inefficient (pg 158).

In stark contrast to something that you cannot pursue (i.e. data that supports the cover of Barron’s saying “the US economy will avoid recession and grow at a healthy 3% pace this year”), this is how Asness taught us to think about math/economics:

“I think that good quant investment managers can really be thought of as financial economists who have codified their beliefs into a repeatable process.” Amen brother. A data driven rate-of-change process is both measurable and repeatable.

Pursuing Data - juggling bull 02.19.2016

Back to the Global Macro Grind

Measuring last week’s macro market message was almost as easy as measuring Germany’s PMI reading slowing to a 13-month low of 50.2 this morning – big macro 2016 losers won last week while winners lost some of their YTD gains.

Quantitatively driven investors can see this most easily by looking at the performance of US Equity Style Factors:

  1. High Beta Stocks ramped +7.9% last week (but are still down -14.1% YTD!)
  2. High Short Interest Stocks squeezed +6.9% last week to -6.4% YTD
  3. Smaller Cap Stocks bounced +6.7% last week (but are still down -9.9% YTD)

*Mean performance of Top Quintile vs. Bottom Quintile of SP500 companies

Yes, bear market bounces can be vicious. That’s why you want to have a repeatable risk management #process that provides you the confidence to cover-low and re-short higher.

Unless we have a 1987 type day, fading market moves (both ways) works. And by the way, with the following immediate-term risk ranges, I wouldn’t rule out what might feel like a modern ‘87 type day either:

  1. SP500 immediate-term risk range = 135 points wide at 1
  2. US Equity Volatility (VIX) immediate-term risk range of 19.92-29.28
  3. Oil (WTI) and Oil Volatility (OVX) immediate-term risk range of $25.99-33.05 and 57-81, respectively

In our multi-duration, multi-factor #process, the immediate-term risk range is what should be considered “probable” (in standard deviation speak). Probable and trending levels of 20-80 volatility are incredibly damaging to anyone levered long the underlying securities that are implying that level of volatility.

In recent history, Ben Bernanke would just come into the macro matrix and crush volatility (and credit spreads) with an unprecedented level of un-elected quantitative easing. But he’s gonzo now and, if anything, you see the Federal Reserve data-mining for reasons to stay relatively tight. They aren’t getting easier. That’s why the US Dollar isn’t collapsing.

As Asness revealed to Eugene Fama (“efficient market theory” professor at the University of Chicago), “the only cure for data mining is an out of sample test” (Efficiently Inefficient, pg 159). So please don’t be a data miner. Stick with rate-of-change and the longest dated time-series you can confirm as good data.

This is why I think the future of finance is already here.

While I am still uber bearish on High Beta, Highly Levered, Small Cap US stocks right now, I am probably the most bullish man on Wall Street when it comes to building a better way. Our profession is so far behind the curve on pursuing best fundamental and quantitative research practices that we can only get better from here at an accelerating rate.

One way we’ve really improved our #process since calling the last US stock market crash has been getting rid of broken Old Wall “sentiment readings” and building our own. On that score one of the most important #behavioral read-throughs in an oversupplied hedge fund industry is weekly CFTC non-commercial futures and options positioning. Here’s last week’s:

  1. SP500 (Index + Emini) net SHORT positioning got less short by 60,883 contracts to -162,655
  2. Long-term Treasury (10yr) net LONG positioning only got 11,613 longer last week to +3,599 contracts
  3. Gold net LONG positioning ramped +21,022 longer last week to +93,934 contracts

In rate-of-change terms, how I read points 1-3 are as follows:

  1. After shorting the YTD lows, hedge funds were forced to cover higher last week
  2. After not being bullish on the Long Bond coming into 2016, consensus is nowhere near bullish enough
  3. After an epic +17% ramp to start 2016, chart chasers got too long of Gold at the highs

So, I’ll wait for 1945’ish on the SP500 to short more. I’ll signal buy more TLT closer to 1.85% on the 10yr UST Yield. And I’ll probably buy Gold again on a pullback towards $1151.  I’ll keep pursuing that positioning until the data doesn’t support it.

Our immediate-term Global Macro Risk Ranges are now:

UST 10yr Yield 1.62-1.84%

SPX 1

DAX 8

VIX 19.92-29.28
USD 95.49-97.91

Gold 1175-1251

Best of luck out there this week,

KM

Keith R. McCullough
Chief Executive Officer

Pursuing Data - 02.22.16 Chart