Thursday, October 21, 2010
Reverse engineering ore prices
A couple of excellent pieces are available today on the looming iron ore glut. The first is by Malcolm Maiden of The Age on Australian miners and their plans to expand production. The second is by Australasian Investment Review via the IBT on broader statistics in the global seaborne trade. Both are worth your time.
At some point this blog will take on the data storm around iron ore supply and demand but that will have to wait. For today, it would like to address the enduring mystery that is benchmark iron ore pricing.
Reliable sources in the industry have confirmed for this blogger that the Q3 (July-Aug) quarterly (benchmark) contract for Rio iron ore is in the high $150s per tonne and BHP close to that.
We also know that the Q3 benchmark was a 23% rise over Q2 (April-June).
We also know that the Q4 benchmark has been agreed at 10% lower.
We also know that the majors use a trailing formula of the preceding quarter's spot prices to calculate the following quarter's pricing.
Having averaged the daily spot market price for Q2 (the things this blogger does!), it has determined that the mean price over the three months is a close enough fit with the actual price changes to approximate the major's formula.
But, when applied to the change in the Q4 benchmark this no longer held. If the daily average of spot prices over the three months in Q3 was the key factor then the price for the Q4 benchmark should have fallen 16%. It actually fell 10%.
Assuming the Bloomberg data used is correct, this suggests several intriguing possibilities: First, the ore price formula is not simply the average price over the preceding quarter. Second, whatever the formula is, prices are more sticky going down than they are going up. Third, customers are going to abandon this system the first chance they get.