More so than at any time since the first Gulf War, forecasting a commodity price has become particularly tricky. Ronald McGrainor, a senior vice president at the Hudson, Ohio, office of futures merchant Refco LLC, has developed a formula for copper prices that he believes has a good track record--but it may never have been tested against so many variables.

Speaking to attendees of the Sixth Annual Great Lakes Non-Ferrous Metals Conference, which McGrainor hosted in Cleveland, McGrainor stuck by his previous prediction that copper prices would begin to rise in the third quarter of 2003, reaching as high as 86 cents per pound as the price average in the second quarter of 2004.

McGrainor considers such factors as copper inventory levels, global consumption, overall industrial production numbers and currency exchange rates to help him forecast future copper pricing directions. He tends to focus on long-term trends and patterns, which often feature peaks and troughs that move in six-to-nine year cycles.

Predicting those cycles can be thrown off, said McGrainor, when one independent variable has a prominent effect on the market. One such variable occurs when energy prices spike, he noted.

"Historically, [energy price] spikes are followed by lulls in business spending for the next three to five months," said McGrainor. Such a lull in the U.S. would suppress copper consumption and build up global inventories. He noted that in each of four prior energy price spikes, copper consumption had been building, "but then the energy spike killed it." This time around, consumption had not even begun to escalate before an energy price spike hit.

So, while long-term trends seem to indicate increased demand and better per-pound copper prices, McGrainor says he is "afraid that energy price spike is going to mean some ugly economic information coming out over the next three to four months."