Beginning of the End
Could the commodity and equity rallies have ended? Has a change been triggered by a resurgent dollar, meaning its purportedly strong negative correlation with gold and commodities continues to drive the process? But if so, would this also not signal the end of the reflation and carry trades? Would it not also mean that contractionary monetary forces may soon dominate? If that is the case, is other evidence that the correlation between gold and the dollar, which on average actually has been weak, might be shrinking to a benign level? Looking at the historical record, has this correlation been steady, or has it cycled up and down or faded away for extended periods?
Based upon theory and evidence, the Conservative Economist presents a case that the negative correlation of the dollar and gold has been a weak, on-and-off phenomenon, subject to degradation. This conclusion augments the theory presented in a companion essay published last week, For Whom the Gold Bell Tolls, which illustrates how central bank and fiscal interventions can be overwhelmed by an instinct to hoard. After credit booms, holders of capital distinguish between money and money substitutes, discounting the value of the latter while the former retains or increases in value.
Discounting is a form of inflation, but it differs from inflation associated with monetary expansion seen when leverage is increasing systemically, wherein all forms of money lose value equally, and demand for credit exceeds demand for money. If we are on the downside of a credit boom rather than still amid a continual rise, then we should expect such differentiation between gold and money substitutes to develop, and the correlation between the dollar and gold’s price changes to diminish.
Not being willing to be responsible for poor decisions when managing someone else’s capital, most on Wall Street are evidence-based investors. The most famous was Long-Term Capital Management (LTCM), which back-tested spread data and bet using massive leverage that standard deviations would rarely be exceeded, or not by much. Nassim Taleb shattered the orthodoxy of such risk modelers by discrediting the assumption of the normal distribution, instead preferring Gaussian models, which in nature are more typical. Despite setbacks such as LTCM or more recently, the statistical arbitrage category of hedge funds, money is still largely guided by normal distribution analyses. Such capital therefore is a reservoir that could be directed at new trends once they are apparent.
Austrian economists, of which I am unreservedly one, discount empiricism altogether but rely on theory (backed by historical analysis of human actions). This school of economics finds that fractional reserve lending promotes credit bubbles, which create horrific clusters of error by businessmen, and of course, investors – particularly those reliant upon usually “reliable” empirical data. With due respect to the shortcoming of making assertions based upon short-term data, this discussion nonetheless makes use of correlation data to reveal subtle changes to the tone of trading since the financial crisis walked on the world’s stage in late 2008. These changes might act as guideposts at what may be another inflection point in the market and counter popular attribution of causes for the new trends.
From this Austrian’s theoretical point of view, the bubble today rests not in the gold market, but in the buildup of credit, which in a fiat monetary system happens in tandem with an increase in the broadly measured money supply (M3) during the upswing. M3 includes many types of what an Austrian would call money substitutes, usually paper units created by banks through deposit lending, which begets more deposits. Moreover, since the world’s currencies are pyramided atop the dollar as a reserve and dollars are created through this process, these currencies represent a derivative of a derivative of base money.
What does it mean if the world’s major currencies weaken relative to the dollar? In any minor fluctuation it would merely indicate a global preference of businessmen to invest in the United States, or in dollar-denominated projects. When the internet craze was sizzling, money was drawn to Silicon Valley, which once required dollars when technology talent and manufacturing resided close to corporate headquarters. Now that signs of deleveraging have been abundant since 2008, the interpretation is that a rising dollar signals more demand for money itself, with electronic Federal Reserve Notes acting as an imposter for gold. Gold was once at the center of all things monetary, as confirmed by tales such as the Celtic leader Brennus demanding his weight in gold be paid by the inhabitants of Rome in exchange for being left to live after the city’s sacking in 390 B.C, and its role was usurped only in recent times.
Armed only with theory, an Austrian economist doesn’t know precisely when or if we have reached a turning point. Sometimes such flashes of insight are apparent solely from the preposterousness of contemporaneous explanations of economic happenings. And, in For Whom the Gold Bell Tolls, I explain that the third (and most powerful) phase of the gold market might begin once correlation-driven investors see the dollar’s influence on gold wane.
Lack of clarity is a hallmark of inflection points, until after the fact, when fundamentals seem obvious. Almost overnight a strong consensus behind a falling dollar faded, and in contention for the next trend are a variety of opinions for the economic future: inflation, deflation, V-shaped recovery, double-dip. In the U.S. and Europe, where banks are “all loaned up,” hoarding could generate demand for traditional, debt-free money (gold) in the face of deflating asset prices – particularly for real estate. But symptomatic of an earlier stage of fiat lending excess, developing nations such as those in the BRIC region have produced robust credit growth in the aftermath of the 2008 crisis, rekindling traditional inflationary depreciation of the value of money there. Monetarily one side wants to contract; the other expands, but in both cases governments have opened up the throttle.
When two weather fronts collide, who knows what kind of storm might develop. Quite a bit of stored energy rests within Wall Street’s correlation models, so let’s get a weather report:
The “X” pattern in Figure 1 below illustrates the falling dollar (in white) and the rising gold price (in orange) evident over the past seven years. If you are a trader and you think the dollar has bottomed, you rejoice in the news of deteriorating credit among the PIIG countries (Portugal, Ireland, Italy, & Greece), and bet against gold accordingly.
Figure 1: Trade Weighted Dollar & Gold, 1990-2010. Courtesy Bloomberg.
But in Figure 2 below one has to wonder about the strength of the inverse relationship, and whether it is stable over very long time periods. The correlation is weak to begin with, with an R2 (R2 is the correlation squared) being just 0.30 from 2002-2009, roughly when the “X” pattern was evident. But from the beginning of 2009 to the present, the correlation has deteriorated, producing an R2 of only 0.13.
The empirical data suggest an already weak causality may be slipping close to nil, and eyeballing the general trends of gold and the dollar prior to the internet bubble makes one wonder if we might drift back into a long period when other determinants of the gold price could surface that are more important.
Figure 2: Trade Weighted Dollar & Gold Regressions, 2002-2009 & 2009-2010. Courtesy Bloomberg.
In 2008, on the margin capital flowed in an increasingly bearish direction, and eventually the system collapsed. Recent willingness to bet on a strong dollar and weaker gold prices seems to repeat this sentiment. The interest-carry and reflation trades were strongly in vogue as long as the dollar was in freefall. Now that the dollar might rise, it is logical that base metal prices would be at risk.
But what if gold is detaching from the relationship, as suggested by the breakdown in correlation that began in 2009? What if during 2009 some of the correlation money had gone long base metals and equities, but carved out gold as less appealing in a recovery? With this lessening correlation and an emerging characteristic of being a safe haven against an unpredictable end result of unprecedented monetary and fiscal stimulus, whether it be inflation, deflation, or an inflationary depression, such realignment is logical. Although gold hit an air pocket in late 2008 during the credit crisis, it was the only asset to hit new highs after the dust settled, and oddly it was preferred by those who anticipated a deflationary collapse in the banking system or hyperinflation – opposite ends of the spectrum!
Recently some have argued that the credit problems of the PIIGS are inconsequential, because their economies when combined compare in size to a modest American state. This observation is just as true as the explanation that the sub-prime mortgage crisis would be contained because it was not large. However, both thoughts confuse cause and effect. Small troubled sectors don’t cause unleveraged systems to implode. They are symptoms of a credit bubble beginning to unwind, which depending upon how large it had been inflated to begin with, would make those trifles just be the first of many to fail.
An outcome of gold regaining its sea legs and sprinting to new highs could be consistent with acceptance of drastically lower collateral values, which could not be papered over with idle bank reserves or the endorsement of historical cost over mark-to-market accounting.
Figure 3: Trade Weighted Dollar and Gold Correlation Time Series 2000-2010. Courtesy Bloomberg.
What could trigger this scenario? Again, the movement of correlation-driven money could. Notice in Figure 3 above that while on average correlation between gold and the dollar might be weak, in reality the average is made up of periods of strong negative correlation punctuated by sharp breakdowns when the relationship actually went into positive correlation, or mere spikes within negative territory. Note, too, that key moments like the green spike in late 2005, early 2009, and the three spikes in 2001-2002 preceded sharp rallies in gold. These moments undoubtedly are ones in which investors were forced to acknowledge new realities about the economy, financial markets, and gold’s role in it.
It may be premature to say that the correlation may be headed into another green spike. It may even end its upward thrust somewhere below the zero correlation line. But either way, the reduction in correlation that started in 2010 in Figure 3 or the lower R2 evident in the regressions from Figure 2 might be telling us that gold’s uniquely hitting new highs after the meltdown of 2008 may signal not a bubble in gold, but a spreading awareness that the real bubble is in credit, be it privately created or born from Keynesian stimulus. And if that is the case, this conclusion is so far from the mainstream that the massive amount of energy stored in correlation-based investment pools could dramatically recalibrate the relationship between the relatively minute gold market and the considerably larger stock of fiat currency.
[For more of William W. Baker’s financial commentary you can visit his website here: The Conservative Economist.]