This week an ETF tracking the price of copper, iPath DJ-UBS Copper TR Sub-Idx ETN (NYSEARCA:JJC), saw its 50 day moving average cross below its 200 day moving average. In non-technical terms, copper prices seem to be losing upward momentum.
Click to enlarge:
Click to enlarge
Apart from the fact that our extensive research indicates it’s generally a good idea to exit an ETF when the 50-200 downward cross occurs, there is another reason to be concerned about copper’s price weakness. There is an oft-repeated saying that “copper has a PhD in economics”. Copper is used in many industrial applications and because of that it is alleged to be a great bellwether indicator of economic activity. Furthermore, many market pundits believe that copper is actually a “leading indicator” of GDP or of the stock market. Here is what one such believer recently wrote about the metal…
Copper has broken its uptrend line. Now it may even be forming a bearish “head & shoulders” pattern. If it closes below the black “neck line”, then the global economy is likely heading south… Long before most economic indicators will pick up on it (since most of them are lagging indicators).
This quote comes from a site labeled “Secrets of the Forex Market”. The implication given is that traders would have a fantastic edge on other investors if they would follow the direction of copper prices. We prefer to make investment decisions based on hard scientific evidence rather than hearsay or conventional wisdom, so we set out to evaluate copper’s predictive ability. We were able to gather monthly copper prices back to 1980 from the IMF’s website. We then put this data side by side with quarterly U.S. GDP levels (also available from the IMF) and monthly S&P 500 data, and started to run a few correlations. We converted the GDP quarterly data to monthly data by prorating the quarterly data across the non-quarter end months.
In summary, copper turned out to be a very good coincident indicator of both GDP and the S&P 500. Copper prices had a .50 correlation with the S&P 500 level and an even closer .70 correlation with GDP. These are relatively high numbers. However, we could find no evidence whatsoever that copper had any predictive ability in relation to the stock market or economic growth. When we looked at correlations of the current copper price with GDP and S&P 500 levels 3 months, 6 months, 9 months and 12 months later, the correlations were no higher. This would imply that copper had no predictive power by these measures.
We also took the change in copper prices for the last 1 month and last 3 months and looked at the correlation between those changes and subsequent S&P performance and GDP changes. Correlations were slightly higher here but were only 0.16 at most, not enough to be considered statistically significant. The chart below summarizes our findings.
Click to enlarge:
It is possible that we haven’t tested the correct time-lag variables and it is also possible that copper may have been a good indicator prior to 1980. So if any of our readers have evidence of copper as a good economic or stock market predictor we would love to hear from you. Until then, we are forced to assume that copper’s “PhD in Economics” is of no use in bestowing predictive ability. Then again, maybe the description is appropriate - economists have a horrible track record in predicting the economy or the stock market – just ask the Fed Chairman. But that’s another article entirely…
Disclosure: I am long RJI.