Why Record Stock Correlations Are An Adverse Feedback Loop To Market Participation
The ongoing retail abandonment of stocks is well into record territory, which may occur for any of a variety of reasons (need for cash via redemptions, focus on return of capital than on and shift into fixed income, market distrust, etc.), yet the resulting increasing relative participation by electronic feedback loop chasers and by beta levered players, and the subsequent increase in implied and absolute stock correlations may be a vicious circle that will make future retail participation increasingly more difficult. We have been warning about the threat of lack of stock diversification for many months, although have not been able to explain it as succinctly as BNY's Nicholas Colas succeeds in a recent note, titled "Looking for diversification in all the wrong places", in which he concludes: "If you cannot use diversification to manage incremental risk, then why would you take on risk in the first place?" This hits the nail on the head in terms of the ongoing and future lack of stock inflows, since in simple terms in makes the ever greater correlations between various asset classes a barrier to entry for all those very rational investors who seek to diversify bullish or bearish bets with a matched trade. In other words, the market is receptive only to those who blindly wish to bet it all on black or red (with or without leverage). And with ever more people chasing ultra short term return horizons (think numerous underperforming hedge funds that only have one month left to generate some returns in Q3 before their LPs send in the redemption notices), and placing all their bets on just one side of the line, those who wish to pursue rational trades, and generic long-short funds, increasingly obsolete and redundant. All in all, this is a perfect storm for a feedback loop that selects only for those who are willing to bet on an ever more one-sided, and thus unstable, market. Have we gotten to the point where only another market crash will provide retail with suitable speculative entry points?
Much more from Nicholas Colas on this fascinating topic:
Our latest monthly review of asset price correlations finds little changed in terms of equities: average correlations are a minimal 5 basis points higher than last month and continue to be stubbornly high. Precious metal correlations to stocks were little changed, with de minimis movement between asset classes, while foreign and emerging market correlations remained near historical highs. In terms of fixed income, high yield corporate bonds saw a dramatically lower correlation to equities, while the commonality of investment grade debt to equities spiked. Of the currencies we track – Aussie Dollar, Yen and Euro – only the Euro displayed significant change, increasing 30 percentage points in its correlation to U.S. equities. The bright spots here are that gold and silver are doing yeoman’s work providing much-needed diversification to investors who own them, and the Euro’s behavior suggests that the market is less worried about sovereign debt risk hitting that region in the near term. As a larger point, we still find that market correlations are a real impediment to a sustained rally in risk assets. If you cannot use diversification to manage incremental risk, then why would you take on risk in the first place?
Whether you want to admit it or not, you’ve likely done it at least a few times; you’ve put your own name into the Google search box and pushed “Enter.” Was it vanity? Or curiosity? Or did you realize that your employer/spouse/friend probably does it anyway and you should know what they see?
Regardless of motivation, unless your name is truly unique, your search will result in a very off-putting realization: lots of strangers have your name. Like your exact name. Most people, we suspect, could have a pretty good sized dinner party with people that live within a few hundred miles of their home and share their entire name.
Google “Beth Reed” and you’ll see what I mean.
There are dozens of “me.” Maybe hundreds. All with the same name, but obviously completely different. Beth Reed, social work professor. Beth Reed, college soccer player. Beth Reed, pianist (entertainment for that dinner party, perhaps). Beth Reed, desperate job hunter. Beth Reed, cellist (wonder if she practices with the pianist…).
If you know me, you’ll know that I share very little in common with all the other “Beth Reeds.” Same with my Google images search – lots of Beth Reeds, but none who even come close to resembling me. (Scratch that whole “perfect crime” scenario where I have a long lost twin out there in the world). Actually, as a side note, I was kind of pleased to find the #2 Google search result was, in fact, really me. Granted, it’s a link to my Facebook profile (don’t click it, it’s blocked) and not some incredible accomplishment or noteworthy website. But of all the Beth Reeds on Facebook (and there are plenty), my profile popped up right at the top.
If stocks, bonds, currencies and commodities could Google themselves, we bet they would all feel some of the same consternation that I do looking at all these total strangers with MY name. Sure, they all think they are different. Different fundamentals, different cyclical money flows, even different cultures and languages. But many of them really answer to the same name. I’ll explain in the remainder of this note.
Every month we track asset price correlations for a variety of financial assets as well as precious metal commodities. This month shows little break from the recent highly correlated price movements amongst these asset types. There are a few exceptions – we’ll note those in a moment – but by and large a whole range of disparate risk assets continue to trade as one undifferentiated mass. It’s almost as if you typed in “investments to buy” into Google and got pictures of a gold coin, a chemical plant, a 500 Euro note, and the exchange traded fund for U.S. banks. They aren’t the same – not by a long shot – but to the global capital market they all LOOK the same.
That’s not a healthy market. The mathematical benefits of diversification require assets that exhibit low-to-no correlation amongst themselves. When everything moves in synch, asset allocators have to pull in their horns. Wonder why investors are shunning risk and buying bonds? Part of the reason is clearly that the historically proven benefits of diversification just are not working as well as they once did.
U.S. equity correlations among the 10 industrial sectors of the S&P 500 remain near historical highs, as 7 out of the 10 sector ETFs show correlations with the S&P 500 in excess of 90%. Only Healthcare, Utilities and Consumer Staples are lower, and they’re stuck in the 80% range, which is still very high. During August, there were 3 notable moves among the sector ETFS:
- Healthcare stocks’ (XLV) correlation with the broad index fell almost 4 percentage points, while sector returns were down 1.1% over the same time period versus -4.7% for the S&P 500 .
- Consumer Staples (XLP) saw correlations increase more than 5 percentage points, a little surprising although 4 of its top 10 holdings that together make up about 20% of the index are down over the past month (CVS is down 11.43%, for example).
- Correlations of financial shares’ (XLF) increased 4 percentage points, which is understandable since concerns surrounding bank balance sheets and financial regulation are among primary drivers of current stock market weakness.
Precious metals’ correlations to stocks – we track gold and silver – were mixed last month. Silver was basically flat, while the gold correlation to stocks rose 7 percentage points. Both are hovering close to where they theoretically should be – silver (represented by SLV) around 45% (roughly half of silver demand is industrial) and gold (GLD) relatively close to zero at 13%. In short, it appears as though precious metals are among the few asset classes that are doing what they are “supposed” to do – diversify risk versus financial assets.
Foreign and emerging market correlations continue to stick near historic high correlations, with their respective ETFs (EFA and EEM) showing price correlations in excess of 90%.
In fixed income land, high yield bond (HYG) correlations with U.S. equities fell 25 percentage points, largely as a result of stock market weakness over the past couple of weeks versus still decent bond market strength. In contrast, investment grade corporate bonds (LQD) saw their correlation rise almost 30 percentage points.
Of the 3 currencies we track – Aussie Dollar (FXA), Yen (FXY) and Euro (FXE) – the Euro’s 30 percentage point increase in its correlation to U.S. stocks was the only significant move. It would appear that, for better or worse, the fears over a European sovereign debt crisis have receded and been replaced with a more general global macro set of concerns. Hard to read that as really positive or negative at this point.
Lastly, if the S&P sector ETF correlations over the past year are any indication of what’s in store, chances of a significant, lasting market rally appear rather slim. As the last 2 charts show, sector ETF correlations tend to break down and widen when the market is in the midst of a sustained rally, as it did in April and January. Currently, the 10 sector ETFs are bunched together, and though the difference between the high and low correlation has widened slightly, the average correlation has barely budged.
We hope this topic gets much more discussion in the mainstream media, as the implications of these observations are that pretty soon trading stocks, and pretty much any underlying assets will be completely useless, as an investor can get the same risk exposure by buying an ETF of the entire market, such as the SPY. Another tangent is that we have now reached a state of complete tail-wags-the-doggery, where the synthetic determines the prices of its constituent, underlying assets. However, don't tell all the ETF managers out there: if it is discovered that incremental diversification leads to a potential beta delta (yes, a lot of non-insolvent Greek) differential of just 10%, as 90% of beta is now inherent in all assets, this will soon make all vain attempts to pick off beta, let alone alpha, irrelevant. It will also kill all interest in purchasing individual stocks, especially in hedge strategy. We are concerned that this will mean outflows from all non-beta based strategies accelerate in the weeks and months ahead.