It is not the first time USDA surprised markets with an unexpected number, and it’s not the biggest surprise that USDA has ever pulled, but Wednesday’s estimate that corn production will total 13.686 billion bushels was a tough punch in the gut to many hoping for a lower number.
Dow Jones’ pre-report survey found only one out of 23 analysts within 30 million bushels of USDA’s estimate. The average guess of 13.318 bb was 368 mb too low. It was a stunning miss.
There have been many different responses to USDA’s August estimates, ranging from anger to passive surrender. Most analysts I have heard so far seem to go along with USDA’s numbers, knowing that field-based data was included. I would add it is hard to ignore USDA’s survey of 23,000 producers.
While there is no compelling evidence yet to say USDA is wrong, that does not mean USDA is right, either.
If there is anything we should know by now, it is that estimating the U.S. corn harvest in early August is a difficult task. USDA’s own numbers show that over the past 34 years, the 90% confidence interval for estimating U.S. corn production in August is 11.8%. In other words, this year’s corn harvest “should” finish somewhere between 12.1 billion and 15.3 billion bushels. Statistically speaking, that is about all we know at this point.
Of course, that is not how the market sees it. The market has a knack for ascribing certainty to USDA’s numbers — at least in the short-run. On Monday, December corn closed at $3.74 1/2, down 13 cents from last Tuesday’s pre-report close of $3.87 1/2 as noncommercial liquidation likely followed Wednesday’s bearish USDA report. Round 1 went to the bears and it may be a sign of more bearish news to come this fall, but there is no way yet to know for sure. The risk of uncertainty remains high in both directions.
Pro Farmer’s Midwest Crop Tour began Monday and will provide us another piece of the puzzle. The data from the tour does not match the broad scope of USDA’s August survey, but it will give us samples to either confirm or challenge USDA’s findings. With such a large range of variation this year among states and among fields, the more field samples we can obtain, the better.
Some may see this as a tedious exercise of yield-guessing, trying to figure out if corn prices should be 20 cents higher or lower, but there is more at stake. If it turns out USDA’s August estimates are close to accurate, we need to consider the possibility that the path of trendline yields has shifted to a steeper, upward slope.
Consider that in February of this year, Scott Irwin and Darrell Good at the University of Illinois cited a two-in-10 chance of (corn) yield rising above 170 bushels this year.* I understand that 168.8 bushels is not above 170, but the possibility of 168.8 bushels in a year when poor-to-very poor crop ratings for Missouri, Illinois, and Indiana are 17%, 15%, and 25% respectively, suggests that yield potential has increased more than expected.
An interesting comment about yield potential comes from a 2012 study of corn yields by R.L. Nielsen of Purdue University. At the time, he identified two significant periods of improvement in corn productivity. The first was from 1937 to 1955 and the second covered 1955 to the present. Even in 2012, some wondered if a third shift happened “due to the advent of transgenic hybrid technology in the mid-1990s,” but Nielsen concluded, “the data show little evidence…”**
The data was not there in 2012 and it is not confirmed yet in 2015, but the possibility remains that we are entering a new era of increased corn productivity. Just last year, Macon County, Illinois had the highest average yield for that state, at 236.0 bushels an acre.***
After USDA’s national estimate of 168.8 bushels an acre last Wednesday, it is fair to wonder: Is this is a sign of things to come?