ANALYSIS: Trade war, storms highlight flaws in USDA crop data

19 Jun 2019 | Tim Worledge, Andy Allan

Wild swings in corn futures are expected to continue in the weeks ahead, as the trade struggles to interpret key USDA data that normally provides price guidance, market sources have told Agricensus.

With the corn-belt having undergone one of its biggest drenchings since records began, much of the trade has been focussing on regular USDA data for guidance on when and how much corn has been planted as a guide to 2019/20 US production.

However, the USDA’s seminal crop progress report – which on Monday said 92% of seed was in the ground – is increasingly viewed as misleading as different corn-growing states use different methodologies to estimate progress.

“I think we can say with 100% confidence right now that 92% corn planting progress does not mean 92% of the 92.8 million acres from the March Prospective Plantings report,” said Scott Irwin, an agricultural economist at the University of Illinois.


And that’s the nub of the concerns – with corn futures collapsing in March after the USDA revealed expectations that farmers would switch away from soybeans and towards corn – the intervention of near constant rain since has left that forecast in tatters and prompted a major rebound in futures prices.

Thoughts are now turning to what the rain may mean for 2019/20 production estimates, with the effort complicated by the prevalence of prevented plantings – an insurance declaration that guarantees payment for farmers in the event of them being unable to plant in the optimal window.

With some analysts already anticipating the loss of up to 10 million acres versus the March estimate to prevented planting, others are also expecting yields to be hit hard, as incessant rain raises fears that fields already planted may also need to be written off.

Typically, the maths is relatively simple – on any given year, analysts and forecasters take the USDA’s March plantings report – a survey of farmers intentions – and multiply that by the percentage planted each week to give an idea of acreage in the ground.

To get production figures, they then look at weather on a daily basis to figure out a yield.

However, the weeks of heavy rain across the Midwest and a stop-start US-China trade war has introduced huge price volatility in corn and soybean futures, sparking speculation that the prospective planting report was wrong before it was even published.

Black hole

And if that wasn’t a big enough problem, some states when estimating progress include prevented plantings as planted acreage, while others don’t.

Michigan’s crop report can rarely have been more widely read than its June 16 edition, which acknowledged that its “percent of corn planted advanced not by planters, but the decision of many farmers to opt for prevented plantings.”

But not every state is believed to have included the prevent planting figures, making it almost impossible to build the basic assumptions that underpin forward projections.

“Apparently, there is no set policy (for crop progress reports)… we won’t know for sure until January,” Charlie Sernatinger of brokerage ED&F Man told Agricensus about the likely final acreage.

“Here is the problem we face: we will have no idea what the total acreage will be for actual plantings for months,” said Terry Reilly, an analyst at brokerage Futures International.

On June 28 the USDA will publish its June acreage report, but that is still a survey that was conducted earlier this month.

Producers send in actual plantings around July 1 for most of the Midwest.

Trade view

It all amounts to making the 2019/20 market year harvest an exceptionally tricky one to forecast as it has shown up the flaws in the weekly Crop Progress Report.

“The trade is now 85-ish (million acres) on corn and 80-ish (million acres) on beans,” Sernatinger said, with the USDA typically making a 7 million-acre distinction between acres planted and acres harvested, which could take area harvested to 78 million acres.

“Then we get to talk about yield,” he added.

If realised, that would take planted area down by around 10% on initial corn planting estimates and 4% down on soybean projections.

If yields of 166 bu/acre are also realised, then the harvested US corn crop could sink as low as 12.9 billion bushels (329 million mt).