Feb. 10, 2011
It was a difficult newsletter to write this week. After spending last week at the 2012 Top Producer Seminar where I moderated a panel discussion on the volatility created by recent USDA reports, farmer's frustration about the reports has been an often-discussed topic here at Pro Farmer. We asked some Pro Farmer Members and some long-time market- and USDA-watchers for input and consensus was clear: USDA needs to do an internal review of its processes used to provide the critical information the markets require to accurately execute the duty of price discovery.
In an effort to explain our call for USDA to review its processes, here's the content from this week's News page 4.
Yes. Without doubt, USDA’s supply estimates for crops and livestock grown in the U.S. and around the world are relevant. But the simple fact some farmers question the relevancy of USDA’s estimates speaks volumes of how frustrated they’ve become with the "surprise component" of recent reports. One or two surprises... okay. But USDA’s combination of the National Ag Statistics Service (NASS) and World Ag Outlook Board (WAOB) have been pinning the tail on the donkey at this surprise party since June of 2010. That’s too long for many farmers who are losing faith in the numbers.
Data versus a very educated ‘guess’ —
NASS conducts supply-side surveys for USDA: Crop Production; Quarterly Grain Stocks; Cattle Inventory; Cattle on Feed; Quarterly Hogs & Pigs Reports. Those are the headline examples of the surveys NASS conducts.
None of the survey work NASS does is easy and there is a strict process of data gathering and data checks in place that turn out the best set of supply-side data of any country in the world. Without doubt, data from NASS is far more reliable than production estimates from the governments of China, Brazil, Argentina or any other country. But having a high level of reliability does not imply complete accuracy. Nor should 100% accuracy be expected. What we should expect is the highest level of accuracy possible.
WAOB "takes" the NASS supply-side estimates and balances the supply against demand estimates. Yes... these are "guesses" of how much will be used, but it is a very educated guess made by highly experienced and "just plain smart" people.
The overlap of some data is why we refer to supply or Supply & Demand Reports as "USDA Reports," rather than specifying NASS or WAOB.
Understand the origin of ‘surprises’ —
NASS has taken many shots for its Quarterly Grain Stocks Reports. In many reports, starting in June 2010, NASS has delivered grain stocks well above or below trade expectations. In June 2010, NASS numbers indicated March-May 2010 corn use was about 300 million bu. bigger than trade expectations, resulting in June 1, 2010, corn stocks that were 300 million bu. below the average pre-report trade expectations.
It’s how those numbers compare to trade expectations that create the surprise.
Based on June 1 corn stocks and the indicated usage pace, WAOB on Sept. 10, 2010, estimated 2009-10 corn carryover at 1.386 billion bu., setting traders’ expectations for Sept. 1, corn stocks. Instead, NASS on Sept. 30, estimated Sept. 1 corn stocks at 1.7 billion bushels. (It appeared USDA had "found" the bushels that were "lost" in June.)
A similar scenario played out in September 2011. USDA’s NASS indicated Sept. 1, 2011, corn stocks of 1.128 billion bushels. Just days earlier, WAOB had estimated 2010-11 corn carryover of 920 million bushels.
While Sept. 1 corn stocks (from NASS) have been above pre-report trade expectations the past two years, those trade expectations were based largely on carryover estimates (from WAOB) delivered just days earlier.
To be fair, the WAOB carryover estimates are based on stocks estimates and indicated usage trends from NASS delivered earlier in the marketing year.
Yeah... it’s kind of a circle with what appears to be conflicting data flowing from "USDA."
USDA data is a benchmark —
Benchmark data is the data set to which all other data is compared. Simply put... USDA is right because it’s right. It is the best set of data available.
We run the Midwest Crop Tour and gather a large amount of data on corn and soybean yield potential. At the end of the Tour, your editors take that data and use it to estimate national average corn and soybean yields and to estimate crop size. With some notable exceptions, we’ve done a good job of estimating yields and crops. How do we know that? Because we compare our estimates made in August to the final yield estimate delivered by NASS in the January Annual Production Summary. How do we know there are "notable exceptions?" Because we compare it to USDA’s crop estimate. That makes USDA the best crop estimate in the business — the benchmark estimate.
That does not, however, mean we’ll stop doing the Crop Tour and making yield and crop estimates. We’ll also continue to estimate use and explain how we see the marketing year playing out differently than does USDA in the Supply & Demand tables. And we will advise risk-management strategies based on our expectations in anticipation that USDA estimates will (eventually) match up with our expectations.
If our assumptions are wrong, we’ll adjust marketing strategies to account for those differences.
But now is the time for a review —
Regardless of track records, it’s good to internally review processes from time to time to make sure "old, consistent" formats still apply in today’s world. We did that a few years ago for Crop Tour, resulting in an adjustment to the corn yield calculation.
With farmer, industry and commodity groups’ frustration about USDA’s reports, it’s time for a review. A review does not assume the data is wrong or USDA has delivered "bogus" numbers. Years ago, NASS reviewed soybean stocks data and improved the system. It’s time again for a review to see how the system might be improved.