For much of this year macroeconomic forces have propelled price moves in agricultural commodities, as we recently noted, but a new analysis of similar historical periods of dislocation suggests that a return to fundamentals in agricultural commodity markets could now be underway.
Since 2000, the correlation between an equally-weighted basket of agricultural commodities (wheat, soybean meal, soybean oil, corn, and soybeans) and 10-Year US Treasury yields is 0.07. Additionally, the correlation between this basket, created using data from the Gro Platform, and crude oil prices over the same 22-year period has been 0.30.
These correlations peaked at 0.92 for agricultural commodities versus crude oil in early June and at 0.90 for Treasuries in late May. And these correlations have now started to show signs of loosening, suggesting a more normalized environment. The correlation between the Gro agricultural commodities basket and 10-year Treasury yields is now at 0.64 (see chart below).
The correlation between the Gro basket and crude oil has also fallen slightly, but it remains elevated at 0.82. As soybean oil and corn, two items in the Gro basket, are increasingly important feedstocks for renewable fuel, their prices are generally more closely correlated with crude oil prices.
Our analysis also shows that even though the correlation between crude oil and agricultural commodities, like wheat and corn, is still above average, the most recent spike in correlations started to come off in early June — 34 trading days after the trading pattern became established.
By comparison, the average duration of the 12 other instances in which macro forces such as movements in Treasury yields overwhelmed ag fundamentals since 2000 was 87 trading days, or approximately four full months. These other instances, however, include some periods where the correlations only briefly exceeded our one standard deviation from the average threshold, as can be seen in the chart above.
For our analysis, we created an equally weighted basket of agricultural commodities using data from the Gro Platform. We then examined the rolling one-year correlation between this Gro basket, the yield on the 10-year Treasury, and the price of crude oil, noting only instances in which the correlation exceeded the average by one standard deviation or 68% of all data points. This analysis can be replicated in large part with data from the Gro Platform and our Excel Add-in.