UW–Madison Smart Restart: For information about fall semester instruction and campus operations, please visit For COVID-19 news updates, see

During this time, please contact us at

UW-Madison Researchers Stand By Their Conclusions About Manipulation of the National Cheese Exchange

Two UW-Madison agricultural economists haven”t backed off a millimeter from their conclusion that major cheese traders — particularly Kraft Foods — intentionally drove down prices on the National Cheese Exchange (NCE) in order to shift prices in the broader cheese market in their favor.
The economists, Professors Bruce W. Marion and Willard F. Mueller, are unimpressed by a cheese-industry-financed analysis of their findings conducted by University of Maryland economist Bruce L. Gardner. Gardner detected no evidence of manipulation of cheese prices and concluded that trading on the exchange simply reflects cheese supply and demand.
“Gardner”s econometric analysis is seriously flawed. Moreover, he totally ignores the voluminous documentary evidence that, to us and most readers, is even more compelling than our econometric analysis,” Marion and Mueller state in a seven-page response to the Gardner study.
The UW-Madison authors say that Gardner”s review simply “repackages” criticisms Kraft made at Congressional hearings in May after the UW-Madison study was released.
The original 210-page report, authored by Marion, Mueller, Maqbool Sial and Frederick Geithman, was based on a comprehensive, 40-month investigation conducted jointly with the Wisconsin Department of Agriculture, Trade and Consumer Protection. That investigation was triggered by widespread concern about National Cheese Exchange”s heavy influence on prices paid throughout the nation for cheese, and consequently, on prices paid for milk.
The exchange, located in Green Bay, is the only national cash auction market for bulk cheese. Its members — major cheese manufacturers, marketers and brokers — meet weekly for about 30 minutes to buy and sell cheese to each other in carload quantities.
Although little cheese is sold there (less than half of one percent of U.S. production during the period studied) the price of nearly all bulk cheese is pegged to the NCE price.
Marion and Mueller analyzed NCE trading from 1988 through 1993. They found that Kraft — the leading buyer of cheese in the United States — was exclusively a seller on the NCE. At times during this period, Kraft sold on the exchange at prices that were less than it could have received off the exchange.
“This raises the question whether companies may ”trade against interest” on the exchange in order to affect prices to their advantage in the larger cheese market,” they state.
Gardner criticized the UW-Madison study for two reasons. He chided Marion and Mueller for using monthly data to analyze trading that occurs weekly. He also said that Marion and Mueller used an improper measure of individual traders” activity on the exchange.
In response to the first charge, the UW-Madison economists note that whereas weekly data are available for NCE activity, only monthly data are available for supply and demand. You can accurately lump weekly figures together to create monthly data, but doing the reverse introduces errors, they argue.
Furthermore, the UW-Madison researchers say that if Gardner and Kraft consultants had been “responsible econometricians” they would have tried doing it both ways — plugging both monthly and weekly data into their econometric model — and reported the results. The UW-Madison researchers did so with their own model, and got the same results both ways.
Gardner”s other criticism deals with how Marion and Mueller measured traders” activity on the exchange. The UW-Madison model considered traders to be active if they were making bids or offers, regardless of whether those offers or bids resulted in sales. Ninety percent of NCE price changes result from unfilled bids and uncovered offers, they point out.
Gardner”s final conclusion that the NCE prices accurately reflect supply and demand “requires the large leap in faith that not only do traders use the NCE primarily to deal with surpluses and shortages, a conclusion we challenge, but that the quantities offered and sought on the NCE accurately reflect the aggregate industry supply and demand. Gardner has not one shred of evidence to support either point,” Marion and Mueller point out.
“We believe that there is compelling evidence that much trading on the NCE is strategic, entered into to influence prices. For this reason, we also are convinced that the ”shortages” and ”surpluses” that Gardner perceives are not accurate reflections of the industry supply and demand position. Rather, these represent the efforts of a small number of firms to set national cheese prices that represent their interests.”