I don't know if market guru Dennis Gartman speaks French, but his recent announcement of a course correction making him once again “wild-eyed bullish” on stocks reminded me of a classic Gallic expression: “plus ça change, plus c'est la même chose.” The more things change, the more they stay the same.
And yet Gartman, a much-seen talking head on CNBC and publisher of an eponymous newsletter, was also long on gold at the same time. Curious. Aren't gold and stocks portfolio antagonists? Doesn't holding one asset necessarily cancel out the other?
Perhaps you've noticed the increasing correlation between gold and stocks. If nothing else, you've probably witnessed some handwringing about the spike in asset class correlations lately.
If it's been a while since your last statistics brush-up, perfect correlation is represented by a coefficient of 1 (1.00), meaning Asset X moves in lockstep with Asset Y. If X moves up, so does Y. A correlation of -1 (-1.00) depicts a perfectly negative relationship, meaning the two assets move in diametrically opposite directions. When Asset X moves up, then Asset Z moves down. Correlations normally range somewhere between these extremes.
Despite the recent press, the much-ballyhooed convergence phenomenon isn't a monolith. Yes, correlations among equities worldwide have increased in the wake of the 2008-2009 market meltdown, but the relationship of stocks to other asset classes has tracked a completely different course.
Let's use the market sell-off as a baseline. Domestic and non-U.S. stocks, together with real estate equities, pretty much moved in lockstep, averaging a 0.95 correlation coefficient vs. the S&P 500. There wasn't much hiding room for portfolios diversified across these assets. As illustrated in Chart 1, you had a better shot at dampening portfolio volatility if you owned bonds, commodities or gold. Gold, in fact, provided the singular, albeit slight, negative correlation to stocks.
Keep in mind that the coefficients depicted in Chart 1 are snapshots. It would be misleading to characterize bullion's relationship to equities as constantly inverse. Over the past five years, in fact, the correlation between the S&P and gold has wobbled considerably on either side of zero. Look at the rolling 30-day correlations in Chart 2. More than half — specifically, 54 percent — of the 1,200 daily values are non-negative. To boot, the correlation trend — represented by the dashed black line — continues to skew positive.
So what's happened to the correlations of the other asset classes since the sell-off? Not much for equities; correlations have remained high. But if we track the meltdown in six-month segments, we can easily spot significant changes in the values for the other assets.
As you can see in Table 1, correlations have, indeed, become more positive, on average, over the past two years. It hasn't been a global trend, though. Vanilla bond correlations have actually become more negative. If we look at the more universally adaptable taxable bond exposures through the same 30-day prism we used to regard gold, a remarkable image appears.
In Chart 3, you'll note that fully 93 percent of the daily correlation ticks for taxable bonds were negative and, in contrast to gold, the trend skews south.
So what can we take away from this? Has gold's luster as a portfolio diversifier been permanently dulled? Are bonds the risk buster of the future? The answer to both questions is a qualified “no.”
Why? Well, first of all, historic correlation data often mask current market conditions. Most advisors and investors inclined to use portfolio-building statistics tend to rely upon multi-year correlation data. Morningstar, for example, typically cranks out stats based on a portfolio's trailing 36-month performance. As we've seen, market dynamics can be quite fluid. A lot can change in three years. One value encapsulating the entire period may very well obscure trend shifts.
More importantly, regarding asset class correlations alone necessarily ignores the underlying economy. The relationship between asset classes doesn't change in a vacuum. These shifts occur in response to variations in the financial terrain.
The basic economic scenarios of inflation, deflation, recession and recovery are still the primary drivers of asset prices. Nothing new in that. Plus ça change, plus c'est la même chose.
Deflation, of course, makes bond investments more attractive while inflationary fears tend to increase gold demand. Equities are bolstered by rosy economic outlooks. All this is the stuff of a freshman economics course. Or the Series 7 study guide. Advisors, and hopefully investors, have long known of these cause-and-effect relationships.
It seems as if Dennis Gartman now thinks the domestic economy's on the rebound (long stocks), subject to a reflation in asset values (long gold). Of course, the withering of balance sheets over the past couple of years has given them little room to move but to inflate.
Then again, Gartman may simply appreciate the ameliorative effect of spreading his investments over classes that respond differentially to shifts in the economic landscape. There's that old idea again. Build a portfolio of more or less “permanent” assets and you're girded for nearly every storm. It's not that you won't suffer, you're just likely to suffer less.
Take a look at the track record of our asset collection over the past five years in Chart 4 to see this in action.
As you can see, equities have been pretty inefficient assets over the past five years. Collectively, large-cap stocks have struggled just to rise above their pre-meltdown level. The compound annual growth rate for the S&P 500 has been only 1.7 percent. Gold, on the other hand, has been a — ahem — sterling asset. Bullion has risen at a 21.9 percent average annual rate. But gold's volatility — at 22.3 percent per annum — has tarnished its performance. Just ask Dennis Gartman. He's regularly jumped in and out of bullion as it gyrated.
No, bonds have been the darlings of this half-decade. Take the aggregate bond index as an example. True, it's only risen at a relatively modest 6.4 percent annual rate, but its even more modest 4.4 percent annualized standard deviation has made taxable fixed income a much more efficient risk.
Would foreknowledge of that have justified shunning gold as a portfolio diversifier? When you look at the differing results obtained by overlaying an S&P 500 position with a bond or gold exposure, you wouldn't think so. Over the past five years, a 20 percent commitment — rebalanced monthly — to gold would have improved a stock portfolio's performance more than a similar allocation to taxable bonds. Gold boosts the average annual growth rate to 6.1 percent vs. 2 percent for a bond overlay, but at a cost. Portfolio volatility declines more with the bond allocation.
Now look what happens when you overlay both assets on the stock portfolio. Chart 5 traces the allocation results. Not only is the average annual return improved to 7 percent when one-fifth of portfolio assets are committed to both classes, but portfolio volatility is also compressed more than it would be with a single addition — a win/win for portfolio efficiency (see table 2).
This data ought to relax some of the handwringing of investors and advisors who fret over reports of increasing correlations. Yes, gold's correlation to stocks is skewing positive. For now. Bonds are also skewing, but in a negative direction. For now.
It's pretty well established that a portfolio that's properly allocated to assets responsive to changes in the economic environment is more likely to survive to its ultimate liquidation or generational transfer. Interest rates may change; gold prices may vacillate and stocks may wobble, but certain portfolio construction tenets remain immutable.
It seems the more things change, the more they stay the same. Merci, Monsieur Gartman.