Gold & Oil: A Historical Ratio Turns Bad

Tram Footings (circa 1912) - Windfall Mine, Eureka, Nevada
February 29, 2016

The 40-year average

In September 2010, Barrick Gold ex-CEO Aaron Regent cited historical gold ratios when asked if $1,300 was an excessive price for an ounce of gold. During a CNBC Business News interview, he argued that gold was in in line with the 40-year average of the gold-to-oil price ratio or 15 barrels per ounce. Mr. Regent presented a similar ratio comparison for copper. So no, gold prices were not too high.

Last Friday Comex gold at closed $1,220.4 per ounce; and Nymex WTI oil, at $32.78 per barrel. The associated gold-to-oil ratio is 37 barrels per ounce - quite a bit higher than Mr. Regent’s historical average. In fact, the same ratio peaked at a jaw-dropping 47 earlier this month. What’s up?

The following two case studies illustrate the good, the bad and the ugly of historically-based gold ratio comparisons and a hint of what’s to come.

Case I: The U.S. QEs

In fairness, Mr. Regent made his comparisons during a fairly stable period for commodities - after the volatility of the Great Recession and before the beginning of the second round of U.S. Federal Reserve quantitative easing (or QE2). In the fall of 2010, it made good sense that gold ratios should seek equilibrium with long-term averages.

The first phase of quantitative easing (QE1) reflated commodities from the depths of 2008. The next two phases witnessed very different results. Figure 1 is a plot of the gold-to-WTI oil ratio from the start of QE2 to the present.

Figure 1. CASE I: Gold-to-WTI Oil Ratio (October 2010 to the present)

The QE2 program began November 3, 2010, six weeks after the Regent interview. The second phase of accommodative monetary policy weakened the U.S. dollar and inflated dollarized commodities against a backdrop of strong China demand. Oil was further boosted by the turmoil of Arab Spring when oil prices moved above $110 per barrel by April 2011. Oil rose faster than gold causing the gold-to-oil ratio to dip to 13 barrels per ounce that month.

This trend reversed with the end of QE2 in June followed by the U.S. debt crisis that summer. After the U.S. debt downgrade in August, gold soared to its all-time $1,900-plus high by September as oil prices plunged to $80 per barrel territory. Their respective ratio remained above 20 barrels per ounce for most of this two-month period.

The third and final phase of U.S. quantitative easing commenced June 2012 as the global appetite for commodities waned with a slowing China. By that November, gold was still above $1,700 per ounce but would soon begin its 3-year downtrend as investment favored booming equity markets over safe haven. By the end of QE3, oil was again in the $80 range after a long stay above $90; gold was only a few dollars from last Friday’s closing price. The gold-to-oil ratio scored 14.9 barrels per ounce – very close to Mr. Regent’s long-term average.

Had markets returned to calm waters? Hardly.

As QE-3 wound down and the Federal Reserve contemplated future interest rate rises, Japan and Europe poured on their own versions of quantitative easing. This strengthened the U.S. dollar as sluggish demand and global oversupply crushed commodity prices – especially oil. The combination propelled the gold-to-oil ratio to its historic February 11, 2016 peak of 47.1 barrels per ounce. Simple statistics applied to the QE2-QE3 period illustrates the violence of this transition.

The average or mean gold-to-oil ratio is 15.7, again not far from the Regent number. The standard deviation (or sigma) is 2.5 barrels per ounce. Most of the data obediently lie within 2-sigma except for the brief excursion above 20 barrels per ounce during the U.S. debt crisis.

This is another way of saying that the QE2-QE-3 gold ratios fall within the range of what is normally accepted given the common “bell curve” or normal distribution. For this model most data clusters around the center or mean with a few outliers in the “tails” of the curve. The 2-sigma box in Figure 1 suggests this is true with peaks lying outside the box still within an acceptable 3-sigma of the mean. In economic models, a deviation larger than 4-sigma is considered substantial.

By contrast, the February peak is a statistically alarming 12.5-sigma departure from the mean.

Case II: The 1980s Oil Glut & Great Recession

The second case widens the window for analysis to include the Great Recession and the 1980s oil glut. Figure 2 is a plot of the gold-to-oil ratio from the beginning of 1986 to the present.

Figure 2. CASE II: Gold-to-WTI Oil Ratio (January 1986 to the present)

Falling demand following the 1970s energy crisis caused the 1980s surplus of crude, a global oversupply situation similar to today’s. Oil, which had peaked in 1980 at over $35 per barrel, fell in 1986 to below $10. With gold in a range of $350 to $490 per ounce, low oil prices rocketed the gold-to-WTI ratio above 30 barrels per ounce multiple times during 1986-1988.

These high levels were not repeated even during the market volatility of the Great Recession. Then the peak gold-to-oil ratio reached 28 barrels per ounce in February 2009. Prior to this recession, the ratio plumbed 6 barrels per ounce August 2005. Even though the high/low extremes of the extended 1986-QE3 period are greater than the first case, the average over this period is only modestly higher than for either QE2-QE3 (16.0 barrels per ounce versus 15.7) or Regent’s long-term average (15 barrels per ounce).

The Case II extremes do more than double standard deviation of the first case (5.5 barrels per ounce versus 2.5). In this context, the February peak is nonetheless a formidable 5.7-sigma deviation.

Beware the “tail event”

Gold ratios offer a method for determining the value of a commodity relative to gold – an ancient and widely accepted measure of wealth. Since the ratio removes the denominating currency, comparisons over time need not be inflation-adjusted. Gold ratios are also typically expected to revert to the mean over time. This gives credence to Mr. Regent’s argument that gold price in U.S. dollar’s was not excessive in the fall of 2010 because several gold/commodity ratios approached 40-year averages.

The dramatic divergence of the gold-to-oil ratio after October 2014 indicates that something beyond normal market forces is now at work. Even though cases I and II include a witches’ brew of recession, oil gluts and loose monetary policy, the ratio extremes fall within acceptable statistical limits (i.e. less than 4-sigma).

Accordingly, the February peak is a “tail event” -a low-probability, high-consequence occurrence that belies historical expectations. The culprit is the U.S. large scale practice of fracking shale - a technological disruption that has dramatically changed oil's valuation relative to gold. This innovation and OPEC’s defensive reaction to protect market share has created a world awash in oil and slumping prices.

Although the outcome of this conflict is uncertain, it seriously challenges the wisdom of using past ratio averages for a “historical norm” - somewhat like comparing rubber tires to wooden wheels in the history of transportation. I believe it likely the new steady-state ratio will fall somewhere between the February peak and the market tested ratio of yore.

From the heart of North American Gold country – Cheers!

Richard Baker
The Eureka Miner (http://eurekaminer.blogspot.com/)