March WASDE Preview-Argentine Soybeans

On 8 Feb, USDA reported its estimate for Argentine soybean production at 54 million tonnes (MMT) from 18.5 million harvested hectares, implying a yield of 2.88 tonnes per hectare (t/ha). The USDA will update its estimates when it releases its March WASDE report later today.

The trade estimates for today had an average production of 48.36 MMT, which implies a yield of 2.61 t/ha. The low and high estimates are 43 and 53 MMT respectively.

At Gro Intelligence, our estimate is significantly lower: our projected yield is 2.28 t/ha. At USDA’s harvested area, that implies a total production of 42.2 MMT. As this is fully 22 percent lower than the current USDA level, we feel very secure asserting that USDA will lower its Argentine production estimate at the 8 March WASDE report.

The estimate is derived from our machine-learning model, which uses various satellite derived climate and environmental indicators along such as crop health, land surface temperatures, and rainfall data along with Gro derived crop masks to predict county-level yields. The model then aggregates the county-level numbers upward to arrive at a national yield estimate.

Given that the range of trade estimates includes analysts calling for a production increase, our near-certainty of a significant decrease can help our subscribers with their marketing decisions.

According to a simple price vs. carryout regression, the price impact of this report could be significant: while bean prices have risen sharply since the Feb WASDE, they definitely have not accounted for the bullish scenario that the Gro model outlines.

If we assume that prices currently discount the trade estimate of 48 MMT production, deviations from that value have a theoretical price impact of about 40￠/bushel per MMT of change. So, for example, an estimate of 47 MMT would give us a price change expectation from before the report of +40￠/bushel.

Of course, there are numerous caveats and exceptions to this model, but it helps to point out the potentially very significant price impact of this report.

See link for district level yield display: 

https://app.gro-intelligence.com/#/displays/20137