USDA's Packers and Stockyards Administration has filed a complaint against Excel, Cargill's meat packing division. The complaint alleges that the packer failed to disclose a change it made in the formula it uses to calculate the "percent lean" of hogs. USDA maintains that as a result of the undisclosed formula change, about 1,250 producers selling on a carcass merit system received lower prices.

During a period from late October '97 through May '98, producers were underpaid about $1.84 million or about 91 cents per hog, according to USDA.

Excel intends to obtain information from USDA on the lots of hogs for which there were alleged underpayments with the intent to pay producers the full amount identified in the complaint. But, USDA has refused to accept that action as a settlement and may request an administrative judge to impose a fine above and beyond the contested amount.

Excel says it will vigorously contest the complaint. Excel concedes that in early 1997, it began testing various technologies used to measure the amount of lean meat in individual hog carcasses. Excel uses a Fat-O-Meater, a fiber optic probe, to collect carcass readings. The data is then plugged into an equation to give an estimate of percent lean.

Excel says they found that a calibration equation, developed by Purdue University in 1990 for the National Pork Producers Council, gave a more accurate estimate of percent lean than the manufacturer's equation it was using.

"We can take the readings, then go back and take the lean meat off the hog and see how the estimates compared with the actuality," says Excel's Mark Klein.

Klein says the old Fat-O-Meater formula was about 70% accurate and the new formula a little more than 90% accurate. "We looked at six weeks of hogs and compared the two-formula effect on price," says Klein. On average, the price wasn't changed. He says some prices ended up lower, some higher, some unchanged. "We knew what it would get rid of was the extremes; the hogs that really weren't as lean as the old equation was suggesting and probably hogs that weren't as fat as the old equation was suggesting," says Klein.

The complaint from USDA puts $1.04 million of the underpayments at the Excel plant in Beardstown, IL; $780,000 at the Ottumwa, IA, plant; and $16,000 at the Marshall, MO, plant. The change in the formula was not put in at Marshall until late April, six months after the other two plants.

USDA says only two parties were notified of the change. Hog Inc., a producer marketing co-op headquartered in Greenfield, IL, was told of the change in February. Tyson Foods, the former owner of the Marshall, MO, plant was told before the change was made.

Allan Schinckel, Purdue University, who worked on the equation Excel adopted, says that the best way of measuring if one prediction equation is better than another is "out of sample" statistics, which Excel obviously has access to.

"There probably will be no contesting the issue that it did use a more accurate equation," says Schinckel. "The issue is what kind of warning do you give producers that you're changing."

Schinckel says the incident reinforces the importance that producers need to know how percent lean is calculated so they can fine tune what genetics to use and what weights to sell at.