Make Kitco Your Homepage

Copper, gold & the coronavirus

Commentaries & Views

China 2-rho Divergence Indicator© gauges coronavirus impact on copper & gold prices

February 18, 2020

I introduced a market gauge in my January 13th  Kitco commentary to track progress of U.S./China trade as implementation of Phase I begins and future negotiations proceed.  Then the Wuhan coronavirus appeared in Hubei Province. The commodity market stress that followed caused my China 2-rho Divergence Indicator© to approach levels near those that occurred during the height of U.S./China trade tensions in August and early-September. Recently, this gauge peaked signaling the worst may be over for key commodities and the Chinese yuan – at least for the short-term.

Is it a pandemic? Nobody knows for sure; coronavirus infections are being recounted to the upside in China and the death toll continues to rise. This is an economic and viral shock to the world that is hard to quantify. That uncertainty alone underpins a range for safe-haven gold: a solid floor around $1,500 per ounce and the mercurial $1,600-level above - at least for now. If China's growth falls to just several percent and supply chains are disrupted for months, the upper level for gold could be much higher, perhaps $1,800.

Copper and oil prices could head much lower with collapsing demand from the world’s second largest economy. For the short-term, it appears February 11th was a reversal to the upside for these two embattled commodities. My China indicator should prove a useful guide for what may lie ahead as it did for U.S./China trade negotiations this summer and fall.

2-rho Divergence© Technique

2-rho Divergence© is a technique for developing powerful market indicators. The inputs are market variables that share a historical correlative relationship; the output monitors the divergence of those relationships as they deteriorate in distressed markets. The magnitude d of a 2-rho Divergence Indicator© is a number between zero and twice the square root of two (0<d<2.82). Typically, a low number affirms normal market expectations, mid-range values signal impending departures from the norm, and a measure greater than 1.95 is a red flag of extreme divergence. Importantly, a peak in divergence can confirm or anticipate reversals of related market variables from their tops or bottoms.

China 2-rho Divergence Indicator©

The China 2-rho Divergence Indicator© has three China-sensitive inputs: copper price, gold price and the Chinese yuan. China has led copper demand for years. Even as the Chinese move toward a more consumer-based economy less dependent on industrial metals, copper remains an important global benchmark. China produces and consumes the most gold in the world and its citizens consider the yellow metal an important store-of-wealth. Finally, although tightly controlled by the People’s Bank of China, the yuan has a market component and is a bellwether of their economy.

Figure 1 shows copper, gold and yuan over a five-year period through Friday’s closing prices.

In this chart the Chinese yuan is expressed as CNYUSD, the inverse of the more common USDCNY designation

Figure 1a

Figure 1b

Figure 1. China-sensitive input variables: copper, gold & Chinese yuan

Figure 1a plots Comex copper and the yuan demonstrating a rough positive correlation for the two (in this example, the yuan is expressed as CNYUSD so an increasing value denotes currency strength; a decrease, weakness). On January 10th, the U.S./China Phase I negotiation appeared a done deal and it was signed January 15th. Although there was initial exuberance in global equity markets on the trade resolution, copper prices and yuan strength soon declined on news of the severity and spread of the Wuhan coronavirus (1a, blue arrows).

Figure 1b is a graph of Comex copper and Comex gold. Close inspection shows periods of both positive and negative correlation, the latter case associated with times of market stress (copper down, gold up) or exuberance (copper up, gold down). After January 10th gold prices rise followed by the fall in copper prices (1b, blue arrows).

Statistics support the long-term correlation observations. Persistent positive correlations of copper and the yuan over 5 years occur 58% of the time; and negative, only 8%. A correlation is persistent when its short- and longer-term rolling correlations have the same sign – for this analysis, 1-month and 3-month time periods. By contrast, persistent correlations of copper and gold register 41%; and negative, 21%.

Accordingly, the China Indicator is a scalar, d, that measures how far apart these two sets of correlations come together or move apart. The greater the divergence, the greater the magnitude of d. Since correlations are bounded by +/- 1.0 the theoretically highest divergence in the correlation plane is twice the square root of two or 2.82.

Figure 2 is a 5-year plot of the China Indicator with three noted cases of extreme divergence (d>1.95) and a fourth peak associated with the coronavirus outbreak that falls just short of extreme. The average divergence over this period is 0.78.

Figure 2. China 2-rho Divergence Indicator© (5-years)

As I detailed in my January commentary, the first case spans late-2014 and early-2015 when it appeared China’s economy was slowing much faster than expected. This created a “risk-off” sentiment that caused copper prices to tumble, gold prices to rise and the yuan to weaken. Divergence peaked at 2.2465 on January 30, 2015 (point 1).

The second and third cases occurred this summer and early-fall as U.S/China trade tensions escalated. Similar to the 2015 case, a “risk-off” caution drove gold and copper prices in different directions as the Chinese currency weakened. The first peak (point 2) has a maximum of 2.0144 on August 23, 2019; the second (point 3), a maximum of 2.2080 on September 24, 2019.

As expectations grew for a Phase I deal, the indicator fell rapidly to below average divergence with a bottom on January 10, 2020 (d=0.3012).

Although the first cases of the coronavirus in Hubei province occurred in late-December 2019, its impact on global markets was not felt until later in January when the virus was detected in other countries. With some irony the first death attributed to the virus occurred January 11, pouring the first splash of cold water on Phase I exuberance – at least in the commodity space. Divergence rapidly developed to a peak divergence of 1.8399 on February 11th. Friday’s close was off the peak at 1.7223 and trending down (red arrow). A look at how the relevant correlations developed over this period is instructive to understand the dynamics of the coronavirus on markets and what may come next.

Coronavirus 2-? Divergence©

Figure 3 is a Correlation Map (rho-Map©) of copper correlations with gold and copper correlations with the yuan. The time period is January 10 to February 14, 2019. The X-axis is a short-term Pearson rolling correlation and the Y-axis is a correlation over a longer-term. This analysis uses correlation time periods of 1-month (X) and 3-months (Y). Each data point represents a given market-day when the two correlations are computed from closing prices.

Figure 3. Correlation Map of copper-gold & copper-yuan correlations

The data points in Figure 3 are connected in a time sequence to create a Correlation Trajectory for each correlation set. Divergence of the two trajectories is defined as the distance between trajectory endpoints for a given market-day.

Unlike the copper-yuan correlation, copper-gold accelerates away from its positive historical bias. This is typical for the “risk-off” markets experienced in August-September 2019 (i.e. copper down, gold up). Increasing divergence is a result of such departures as evidenced by the maximum separation of trajectories on February 11th (point 4, dashed magenta arrows of length 1.8399).

Importantly, a second peak could occur if the copper-yuan correlations remain in the upper-right quadrant and the copper-gold correlations move deeper into the bottom-left quadrant of negative persistence. Alternately, if 1-month copper-gold correlations continue to decline in negativity (red arrow), the worst may indeed be over for the near-term (i.e. divergence continues to decrease).

Coronavirus impact on copper, gold prices going forward

The 2-rho Technique© constructs indicators based on two sets of historical correlations. Each set contains correlations of two market variables over two time periods. Departures from the norm offer insight into prevailing market conditions. Extreme departures or divergence often provide leading indication of key market reversals from their tops and bottoms and thereby herald a trend to normalcy.

The recent divergence peak of Figure 2 fell short of extreme levels but did confirm reversals of copper and gold prices on February 11. If a second peak appears due to delayed impacts of the coronavirus on supply chains and the global economy it will likely be at severe levels. Past events of extreme divergence demonstrate the power of the indictor to anticipate bottoms and tops. As explained in my January commentary, over the last five years extreme divergence of the China 2-rho Divergence Indicator© flagged three periods of market stress followed by yuan peaks and copper-to-gold ratio lows. The latter, by the Gundlach copper-gold relationship, can often lead U.S. 10-year Treasury yield bottoms (see Note 1).

I intend to write future commentaries to track both the China and Gundlach copper-gold indicator as we continue down the evolving path of the coronavirus and the longer journey of follow-on U.S./China trade negotiations. Hang on.


Disclaimer: The views expressed in this article are those of the author and may not reflect those of Kitco Metals Inc. The author has made every effort to ensure accuracy of information provided; however, neither Kitco Metals Inc. nor the author can guarantee such accuracy. This article is strictly for informational purposes only. It is not a solicitation to make any exchange in commodities, securities or other financial instruments. Kitco Metals Inc. and the author of this article do not accept culpability for losses and/ or damages arising from the use of this publication.