Independent Monitoring Predicts and Verifies Record Soybean Crush Rates

Soybean
Blog April 30, 2018
  • Timm Decker , Senior Analyst, Data Scientist, Agriculture & Biofuels

With the March release of the National Oilseed Processors Association (NOPA) member crush number, facilities have now reported five straight months of record daily crush rates and two of the highest total crush numbers in history. December 2017 set a record high at 166.3 million bushels (ml bu) crushed and was followed by an astronomical value of 171.9 ml bu in March 2018. By leveraging proprietary, near-real-time physical monitoring, Genscape composes its national outlooks and outage reports based on observations from key soybean processing facilities in order to anticipate any market moving numbers such as those of the last five months.

NOPA Monthly Totals Year-over-Year

The string of high numbers was not totally unexpected by the industry as there were several known factors in play that would signal higher numbers. The October through March timespan is typically when facilities crush the most beans. In addition to this, higher soybean stocks and significantly higher crush margins, result in an environment conducive to high crushing rates. Industry analysts responded by predicting relatively high numbers for each of the months to correctly call the increase. What they did not get right, however, was the magnitude.

Compiling the industry average crush prediction, analysts were off by an average of 3 million bushels over this five-month period. Despite the evidence, analysts were hesitant to predict these record setting numbers. This would especially hold true for the all-time high March 2018 number. Prior to this release, the previous seven highest crush totals occurred in October, November, and December making the March record breaker an anomaly. The median estimate for March 2018 was 167.5 ml bu; a full 4.3 ml bu less than the actual number.

Genscape’s Soybean Processing Monitor, which uses independent monitoring and proprietary algorithms to calculate and record near real-time processing rates and outages at major soybean crushing facilities in the U.S., provided clients during this five-month span with estimates within 1.5 ml bu of the true number at an R^2 value of 0.98. During three of the months, the estimate was within 1 ml bu, including the record setting March 2018, where Genscape’s estimate provided to clients on April 3 was at 171.3 ml bu compared to the actual 171.8 ml bu.

Genscape vs. NOPA Estimate

During this period, Genscape monitors detected not only higher processing rates relative to other years, but also a significant decrease in outages, both planned and unplanned. This especially rang true in March where typically we see the first plants taking planned maintenance outages. Using our unique “on-site” monitoring, we were able to keep real-time tracking of facility processing and were able to quantify both the extreme crushing rates and lack of maintenance outages. Genscape’s proprietary algorithm processed this data and returned the record crush numbers completely independent of market conditions or the other subjective clues.

Infrared Genscape Imagery

Heading into maintenance downtime season and with uncertainty regarding possible tariffs on U.S. soybean imports by China, the ability to track U.S. soybean crush rates will become even more important. The Genscape Soybean Processing Monitor will continue to provide clients with daily alerts of facility shutdowns and startups based off real-time monitoring. It will also continue to provide the earliest and best estimates of crushing rates based on what is happening at the plants and free of any emotion. Genscape’s proven predictions of upcoming crush numbers begin in the middle of the current month, allowing for plenty of lead time to prepare for market moving numbers, including the April 2018 estimate which is due for release on May 2.

National Monthly Totals

If you are interested in these industry leading estimates, please click here to learn more about Genscape’s Soybean Processing Monitor.