"That may not be the case this year as demand considerations have become extremely important," said Darrel Good. "Even with expectations of a large corn crop, for example, December 2006 corn futures have rallied from the contract low of $2.33 1/2 reached following the August production forecast to settle at $2.71 on Oct. 6 and trade above $2.80 on Oct. 9.

"With widespread agreement that the October soybean forecast will be larger than the September forecast, November 2006 futures increased almost 25 cents since the end of September. Prices are being influenced by strong export sales for both commodities and prospects for a large increase in corn used for ethanol production and a more rapid expansion of soybean oil used for biodiesel."

Some expect, Good added, that thee positive demand developments will result in larger consumption than forecast by the USDA in September.

"To a large extent, the market is suggesting that the crop size is not all that important this year, particularly for corn," he said. "A marketing year low in the cash price of corn likely occurred early last month and prices now are at levels previously not expected until at least the end of the year.

"The cash price of soybeans is also well above the low established in early September, but there is less confidence that the low price for the year has been established."

Good's comments came as he anticipated the USDA's upcoming October Crop Production and the World Agricultural Supply and Demand Estimates reports both to be released on Oct. 12.

For corn and soybeans, the October production forecast has traditionally been viewed as important because of its historical accuracy. Over the past 36 years, the average difference between the October production forecast and the production estimate released in January following harvest was only 2.4 percent for both crops.

"This compares to a difference of more than 5 percent for the August forecast and more than 4 percent for the September forecast," said Good. "Over the past five years, the average difference between the October forecast and the January estimate was only 1.2 percent for corn and 2.1 percent for soybeans.

"The October report is generally more accurate than the August and September report since some harvest results are reflected in the producer survey and more fruit count and weight data are available in objective yield surveys."

Good said that one reason for the improved accuracy in the October estimates in recent years is that USDA's National Agricultural Statistics Service (NASS) is able to review the Farm Service Agency's (FSA) certified acreage information earlier.

"That information is now available to update planted and harvested acreage estimates in October rather than in January," said Good. "Because of unusually dry weather, NASS also asked producers in Alabama, Georgia, and South Dakota who were surveyed for the October report this year to update their estimate of harvested acreage."

This year, private sources are almost unanimous in expecting a larger production forecast for soybeans in October. That expectation is based on ideas that the yield forecast will be higher than in September.

"Crop conditions ratings showing 62 percent of the crop in good or excellent condition as the growing season ended points to an average U.S. yield of 42.8 bushels, a bushel above the September forecast," said Good. "Yield at that level would result in a forecast of 3.164 billion bushels, 71.5 million above the September forecast if the estimate of harvested acreage is not changed.

"Private sources expect the forecast to be even larger, with the average yield likely exceeding the record 43 bushels of last year."

For corn, there is more diversity of opinion about the October forecast, with most observers apparently expecting only a relatively small change from the September forecast of 11.114 billion bushels, reflecting an average yield of 154.7 bushels.

"Crop condition ratings as the growing season ended showed 61 percent of the crop in good or excellent condition," said Good. "Based on relationships between trend-adjusted yield and the percentage of the crop rated good or excellent at the end of the season since 1986, crop ratings this year point to an average yield of only 148.3 bushels.

"However, the average yield has exceeded that forecast by crop conditions in each of the past seven years. The difference ranged from 0.4 bushels to 7.4 bushels and averaged 3.3 bushels. Expectations about the October yield forecast also vary due to the wide range of yield results being reported."

The market generally does a good job of anticipating the October production forecasts for corn and soybeans. The correlation between the expected change in the production forecast and the actual change in the NASS forecast in October since 1970 is 0.79 for corn and 0.73 for soybeans, with 1.0 reflecting perfect correlation in direction and magnitude of change.

"Even so, surprises in October are not uncommon," Good noted. "The availability of FSA acreage data for the October report increases the opportunities for surprises.

"For example, a 1 percent change in the harvested acreage estimate for corn could change the production forecast by about 110 million bushels and the projection of year-ending stocks by 9 percent."