Seeking Alpha
Commodities, macro, gold & precious metals, long/short equity
Profile| Send Message|
( followers)  

On October 12, the USDA released its latest World Agricultural Supply and Demand Estimate (WASDE). This report showed the acreage of corn being harvested cut by half a million acres (page 12 of pdf). The estimated corn yield per acre was left unchanged at 148.1 bushels per acre.

I found this very intriguing since I wrote an article about forecasting the projected yield which gave a point estimate yield of 143.645. This model had a 95% confidence interval of 147.37 to 139.92. The R^2 of this model was about 90% from 2005-2010. The probability, under normal statistical assumptions, of a value of 148.1, or greater, was .84%.

The fact that the USDA reported a yield estimate that would happen, according to statistics, less then 1% of the time led me to do some more digging. I looked into the estimated yield of the individual states. I compared the condition of the corn crop in the good/excellent category to that of the estimated yield.

Crop Condition Good/Excellent
State: 2010 2011 Change
Colorado 75 54 -21
Illinois 53 45 -8
Indiana 57 35 -22
Iowa 70 59 -11
Kansas 59 26 -33
Kentucky 46 55 9
Michigan 75 65 -10
Minnesota 87 57 -30
Missouri 46 30 -16
Nebraska 80 74 -6
North Carolina 35 21 -14
North Dakota 84 59 -25
Ohio 64 53 -11
Pennsylvania 57 49 -8
South Dakota 73 69 -4
Tennessee 47 59 12
Texas 62 10 -52
Wisconsin 84 75 -9
United States 68 53 -15
October Estimated Crop Yields
State: 2010 2011
Colorado 144 127 -17
Illinois 160 159 -1
Indiana 160 145 -15
Iowa 169 169 0
Kansas 130 105 -25
Kentucky 128 139 11
Michigan 156 148 -8
Minnesota 175 165 -10
Missouri 134 115 -19
Nebraska 170 160 -10
North Carolina 90 84 -6
North Dakota 140 121 -19
Ohio 167 154 -13
Pennsylvania 128 109 -19
South Dakota 145 139 -6
Tennessee 117 137 20
Texas 140 112 -28
Wisconsin 162 160 -2
United States 155.8 148.1 -7.7

I found that the estimate of Illinois matches the historical 5 year regression of condition to yield right on the nose. Here is a graph of its historical plot and the regression equation Yield = 132 + .589 * Crop Condition. I do wonder if this model will hold up for this crop year when July was the hottest in 50 years for the United States and this period is during the critical silking process for corn.

I could not, however, come up with a model to fit Iowa's estimated unchanged yield of 169 bushels per acre, despite an 8 point drop in the condition estimate from last year. I tried adding a time trend but did not come up with anything satisfactory. Iowa is projected to have planted over 16.75% of the U.S. major growing area's crop, so any drop in the yield has a much greater impact on production figures, than a state such as Kentucky which produces less than 2%. 14% of the crop was rated either very poor or poor in 2011. In 2010 just 11% of the crop was in the same category.

Did the state's surveyors take a little break, to have some "medicine", they found in the back 40? But don't worry, you can hope that like a certain politician, they didn't inhale.

Look to be a buyer of corn if the yield is again reduced in the next WASDE report released on November 9th. Until then corn is supported by the cut in acreage but farmers were encouraged to sell into any 50+ cent rally from the lows from the ag web sites I read. Expect choppy trade for the next few weeks but be ready in mid November for another rally if the yield is cut again and farmers are done selling most of their unhedged harvest.

Source: USDA Crop Report: Just What Is Growing In The Back 40?