Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning

11/15/16 09:10PM EST

Key Takeaways:

  1. The OCT retail sales print is a good number in isolation and, more broadly, positively impacts our Q4 real GDP growth estimate, which we are now keen to revise higher to the midpoint of our forecast range.
  2. In GIP Model quadrant speak, this implies a 2nd consecutive quarter of #Quad2 and warrants a considerable shake-up of our long/short factor exposure biases.
  3. The noteworthy changes are: A) no longer buying Treasury bonds, Utilities and/or Gold on pullbacks; B) no longer shorting the Financials on strength; and C) adopting a now-favorable view of U.S. Equity beta.

Recall that our predictive tracking algorithm tracks 30 monthly high-frequency data factors (90 per quarter). With the advent of the relevant data we’ve extracted from today’s retail sales report, we now have 13 reported data points for 4Q16E:

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - U.S. Economic Summary Table

It’s early in terms of reported factors, but the entire point of having a predictive tracking algorithm is to use data confirm or challenge our existing biases. Moreover, the data we’ve received thus far by and large forces us to challenge our existing bearish bias on U.S. real GDP growth here in Q4. Under that strict standard, the obvious implication from the strong sequential consumption numbers highlighted below is that our initial +0.5 QoQ SAAR real GDP growth estimate for 4Q16E is materially divergent from what reported reality may ultimately look like.

Specifically, at the current +4.2% QoQ SAAR growth rate for the retail sales control group, goods consumption is highly likely to be a very solid support for the headline real GDP growth rate in 4Q16E. The fact of the matter is that you’d actually have to see some combination of a massive negative revision to the OCT growth rate and/or material weakness in NOV/DEC to challenge the aforementioned conclusion.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Retail Sales Table

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - RETAIL SALES CONTROL GROUP

From a process perspective, it’s important to review why we apply a stochastic overlay to the Bayesian Inference Process we employ to forecast growth and inflation. Specifically, that process affords us the ability to make assumptions regarding where future growth and inflation readings may fall within a range of probable outcomes. And prior to today, our expectations for growth have been anchored at the low end of said forecast ranges due to our bearish bias on the U.S. consumer.

Recall that our anticipation of incremental softness in the domestic consumption economy was the primary driver of our dour outlook for real GDP growth in the face of dramatically receding base effects here in Q4. We detail precisely what factors were driving said anticipation in the second half of our Q3 GDP recap note titled, “U.S. GDP: Welcome to the Depths of the Cycle”. In short, those catalysts are:

  1. A further deceleration in employment and income growth;
  2. A positive inflection in the personal savings rate;
  3. Cycle-peak base effects for personal consumption expenditures (PCE); and
  4. A continued deterioration in consumer credit conditions that broadly weighs on big-ticket purchases.

We don’t blindly buy into the perma-bull marketing narrative that is the “resiliency of the U.S. consumer”. That said, however, the data is what the data is and it currently paints a picture of resiliency – at least for now.

With respect to the aforementioned easing of real GDP base effects here in Q4, it’s important to note that the deceleration in the 2Y average growth rate in the comparative base period from Q3 to Q4 is actually more than 2x the easing recorded from Q2 to Q3 (in absolute terms). In probability terms, that’s extremely supportive of an acceleration in the YoY growth rate of real GDP here in Q4.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - GDP COMPS

Recall that YoY real GDP growth in Q3 surprised our expectations to the upside by accelerating sequentially because we incorrectly anchored on the low end of our model’s forecast range as opposed to accepting receding base effects for what they ultimately were – i.e. a catalyst for a positive inflection in growth. As such, it would be a mistake for us to continue to do so in the face of positive signals from both the reported high-frequency data and financial markets.

Given the confluence of factors outlined above, we believe it’s now appropriate to toggle our 4Q16E real GDP growth forecast back to the midpoint of the forecasted range. This act revises our YoY growth estimate to +1.7% from +1.4% and our headline (i.e. QoQ SAAR) growth estimate to +1.5% from +0.5%.

Importantly, this new forecast puts the U.S. economy in #Quad2 for the second consecutive quarter to finish 2016.

 

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - GIP Model Baseline

 

Recall that we came into the year anticipating this outcome; in January our intermediate-term outlook for the U.S. economy called for a mid-2016 bottom in growth – both in reported data terms and in financial market terms. A critical review of what we got wrong over the past 3-4 months would seem to suggest we made a mistake in back-end loading our bearishness on both reported growth metrics and market expectations, given that both bottomed out mid-year as we initially predicted.

We can see the aforementioned bottom in the relative performance of Utilities vs. Financials throughout Q3, as well as in the relative performance of [largely overvalued] defensive names vs. value stocks given that the rebound in growth expectations is considered to be a "rising tide that lifts all boats". The latter phenomenon might explain a considerable degree of the recent underperformance of FANG stocks as well.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - XLU vs. XLF

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Russell 3000 Value Index vs. MSCI Minimum Volatility

This puts us between a very large rock and a really hard place. Perhaps we’ve already missed 90-95% of the rates backup and the reversion in relative performance of Utilities and Financials associated with two consecutive quarters of #Quad2. Yes, Donald Trump’s election victory was a “yuuge” factor here as well, but we’d be remiss to ignore the fact that the aforementioned market moves started well in advance of his ascendency in the polls.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - 10Y vs. 5Y5Y

In fairness to us, two of the four pro-#Quad2 sectors (REITS and Healthcare) acted like complete dog poop throughout the summertime and into early-November. Moreover, the other two sectors that have historically outperformed during instances of #Quad2 didn’t do a whole heck of a lot in their own respects. As such, it would’ve been almost as big a mistake to adopt a #Quad2 asset allocation back in early July as it was to not do so. I guess the market doesn't pay up for booming soybean exports when consumption growth gets cut in half...

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Information Technology

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Health Care

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - FTSE NAREIT Index

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Industrials

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - S P 500 Sector Performance

Revisionist history aside, the obvious next step in this process is to look ahead to Q1 and the high probability scenario in that quarter is a return to #Quad3. That would seem to suggest the most appropriate course of action is to fade the recent post-election panic selling in bonds and bond proxies, as well as shorting the associated euphoria in Financials stocks.

That said, however, the fact that inflation is likely to accelerate to its cycle-peak in the Q1/Q2 timeframe may limit any downward pressure on rates from a one-off growth hiccup in 1Q17E – which, as previously mentioned, was always the baseline scenario in our U.S. GDP model. Again, the dour outlook for 2H16E we’ve been highlighting in recent months (prior to now) was overwhelmingly driven by our deep dive on the U.S. employment and consumption cycles and how the concomitant progression therein was likely to dramatically reduce the probability of any base effect-driven rebound in real GDP growth. Sometimes going down the rabbit hole leads to better conclusions; sometimes it leads to more dirt. I can taste a faint hint of dirt as type this note…

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - GROWTH

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - CPI COMPS

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - INFLATION

Returning to the aforementioned discussion regarding where interest rates might be headed next, over longer durations growth has proven a more important factor than inflation in determining the direction of interest rates. Lately, however, inflation expectations have become the primary driver; we can see this by how much the correlation between the 10Y Treasury note yield and the 5Y 5Y-forward breakeven rate has tightened since the beginning of 2H16.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - 10Y vs. 5Y5Y Regression

The key takeaway here is that even if the right call is to avoid chasing the recent mania, interest rates may have limited downside from here to the extent inflation expectations continue to rise alongside the reported CPI figures through at least 1Q17E.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - INFLATION MODEL

Another factor that may limit downside in interest rates from here is the simple fact that the structural growth outlook has improved markedly with the election of Donald Trump and his plan to implement fiscal stimulus in the form of infrastructure spending and massive tax cuts. While it remains to be seen whether or not the formerly obstructionist GOP Congress has changed their minds with respect to levering up the sovereign balance sheet (and Trump plans to do just that), it’s clear that the burden of proof has overwhelmingly shifted to the bears with respect the likelihood and efficacy Trump’s pro-growth policy platform.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Trump Loves Debt

Moreover, it’s important to recall that our GDP model was never able to get to a full-blown recession in reported terms. As such, the risk that the market looks though any Q1 economic softness is perhaps a lot higher than we initially anticipated – especially if a DEC rate hike becomes an easy scapegoat.

Thinking about where we are in the cycle from a structural perspective, the fact remains that leverage in the real economy and debt service ratios haven’t increased by rates that have historically perpetuated the kind of broad-based deleveraging that ultimately catalyzes recessions.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - U.S. Household Credit Gap vs. Debt Service Ratio

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - U.S. Corporate Credit Gap vs. Debt Service Ratio

The household sector is especially delevered and is currently experiencing easing, rather than what has traditionally been late-cycle tightening on the debt service ratio front. The corporate sector, on the other hand, has levered up meaningfully and is experiencing tightening on the debt service ratio front, but not to levels that have preceded the prior two recessions (which as far back as we can get the DSR data). This would seem to suggest the ongoing recession in domestic corporate profits is unlikely to spill over into a full-blown recession – especially given that the pace of marginal tightening in the supply and cost of bank credit has petered out in recent quarters.

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - U.S. Sector Level Credit Gap Just Prior to the Onset of Recession

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - Trough to Peak Delta in U.S. Debt Service Ratio by Sector Just Prior to the Onset of Recession

Process Trumps Politics: Using Data (As Opposed to Political Conjecture) To Improve Our Positioning - SLOS HRM Bank Credit Cycle Indicator

All told, the OCT retail sales print is a good number in isolation and, more broadly, positively impacts our Q4 real GDP growth estimate, which we are now keen to revise higher to the midpoint of our forecast range. In GIP Model quadrant speak, this implies a 2nd consecutive quarter of #Quad2 and warrants a considerable shake-up of our long/short factor exposure biases. The noteworthy changes are: A) no longer buying Treasury bonds, Utilities and/or Gold on pullbacks; B) no longer shorting the Financials on strength; and C) adopting a now-favorable view of U.S. Equity beta.

As always, please feel free to reach out with any follow-up questions. Enjoy the rest of your respective evenings,

DD

Darius Dale

Director

© 2020 Hedgeye Risk Management, LLC. The information contained herein is the property of Hedgeye, which reserves all rights thereto. Redistribution of any part of this information is prohibited without the express written consent of Hedgeye. Hedgeye is not responsible for any errors in or omissions to this information, or for any consequences that may result from the use of this information.