Correlations, Q3 GDP, Clarida/Powell
By admin_45 in Blog
Three “Data” topics today:
#1: There are many ways to measure equity market “stress”, but we favor the CBOE VIX Index and its close cousin, equity sector correlations. The tie between the VIX and correlations is mathematical:
- Under “normal” equity market conditions, sector correlations to the S&P 500 usually range between 0.7 – 0.8. That is an r-squared of 49 – 64 percent. In layman’s terms, the typical S&P 500 sector will trade 50 percent with the market and 50 percent on its own fundamentals over an economic/business cycle.
- Uptrending markets tend to see lower average sector correlations as investors take the macro environment for granted and focus on sector/stock fundamentals. Lower correlations allow the market to trade with less overall volatility since some sectors are up while others are down on any given day.
- Market downturns, such as what we saw last month, push sector correlations higher as investors refocus on systematic issues (interest rates in the case of the just-ended month). This increases overall market volatility since everything zigs or zags together.
Here is the average of the 30-day sector correlations for US large cap Tech, Health Care, Consumer Discretionary, Financials, and Communication Services to the S&P 500 over the last 3 years:

Two points on this data:
- That peak you see right in the middle, at 0.985, lines up with the Pandemic Low of March 23rd, 2020. Every sector was moving in lockstep with the market. Other peaks on this chart, albeit at lower levels than 0.99, also correspond to US equity market pullbacks. The 3 most recent were in mid-May 2021 (0.81 average correlation), mid-July (0.82) and the last day of September (0.84).
- We think it is notable that September’s pullback caused slightly higher sector correlations (that 0.84 reading) than the last 2 at 0.81/0.82. That tells us the most recent market action is different from the last 2 periods of uncertainty.
Takeaway: September’s elevated level of US large cap sector correlations is another way to support the same fundamental message of near-term caution we’ve been espousing over the last month. This pullback, driven by lowered earnings and economic growth expectations as well as higher interest rates, merits a more cautious approach than the usual “buy the dip” advice so common during recent market declines.
#2: Although we covered the Atlanta Fed’s GDPNow model just a few days ago, it is worth a quick revisit today because the model’s Q3 estimate just took another hit. As the chart below shows, GDPNow is now only looking for 2.3 percent growth in the just-ended quarter.

As for what’s “gone wrong” since the end of July, when GDPNow was looking for 6.1 percent, the short answer is “everything”. The table below is a bit of an eye chart (hit the link below and go to page 4 to see it more clearly), but almost every major category in the GDP calculation is down from July 30th. Personal consumer expenditures (PCE), residential investment, CapEx … All lower.

Takeaway: putting aside whether the GDPNow model is right about a 2 percent GDP growth rate, the fact that the model keeps cutting its estimate as late-quarter economic data rolls in is a real problem. We may have entered Q3 with a full head of steam (6 percent), but we apparently exited the quarter at something well less than that.
#3: Fed governors are no smarter than regional presidents, apparently. After Friday’s close, Bloomberg reported that Vice Chair Richard Clarida had personally sold a seven-figure amount of a bond fund to buy 2 stock funds on February 27th, 2020. Just one day later Chair Powell, in response to growing financial market unrest, issued a statement saying the Fed was “closely monitoring developments and their implications for the economic outlook” and would “use (its) tools and act as appropriate to support the economy”.
The Fed has said Clarida’s transactions were a “preplanned rebalancing to his accounts” and “were executed prior to his involvement in deliberations on Federal Reserve actions” to respond to the Pandemic Crisis. We have no doubt that is all true and provable.
However, it is another hit to the Fed’s reputation, coming on the heels of similar issues with regional Fed presidents Kaplan and Rosengren. We’re writing this note on Sunday morning, and for all we know Clarida will resign before we can get this note into your inbox on Sunday evening.
The reason Clarida’s personal trading potentially matters to markets is that this is essentially a “third strike” on a topic most people understand and that puts Chair Powell’s renomination at some greater risk. The average American may not know what the Federal Reserve “does” but asked if the central bank’s leadership should be trading their personal portfolios when an institutional response to a global shock is all but inevitable, we are sure their answer would be “absolutely not”.
Takeaway: Chair Powell’s odds of renomination are down to 66 percent on Sunday morning, from +85 percent before the Kaplan/Rosengren news and 75 percent on Thursday. Markets had thought Mr. Powell was a lock for another 4-year term, a welcomed bit of continuity that reduced the possibility of a 2022 policy error or rookie mistake by a new Fed head. That may still come to pass, but it is certainly not assured.
Sources:
Atlanta Fed GDPNow Model: https://www.atlantafed.org/-/media/documents/cqer/researchcq/gdpnow/RealGDPTrackingSlides.pdf
Bloomberg article on Clarida: https://www.bloomberg.com/news/articles/2021-10-01/clarida-traded-into-stocks-on-eve-of-powell-pandemic-statement?sref=0sRwd4fO
PredictIt Odds of Powell renomination: https://www.predictit.org/markets/detail/7398/Whom-will-the-Senate-next-confirm-as-Chair-of-the-Federal-Reserve