In a break with recent trends, government projections for U.S. corn and soybeans earlier this month landed relatively close to market expectations. But the September supply report due on Thursday has not built a calm reputation over the years.
U.S. quarterly grain stocks, set to be published at noon EDT (1600 GMT) on Thursday, will essentially reveal the ending stocks for the 2020-21 U.S. corn and soybean marketing years that concluded on Aug. 31. The trade trusts that the U.S. Department of Agriculture mostly captured these numbers earlier this month.
But that is not very often the case. Analysts think USDA’s statistics branch will show Sept. 1 corn stocks at 1.155 billion bushels, just 32 million below where USDA’s World Board pegged ending stocks on Sept. 10. By percentage, the World Board has come as close in September only a third of the time in the last 15 years.
At 174 million bushels, the trade guess for Sept. 1 soybean stocks is just 1 million above the September carryout projection, and by percentage that has not happened in over two decades.
Corn stocks have become more difficult to predict because revisions to prior quarters became very noticeable the last couple years. The largest one came last September, when June 1, 2020, corn stocks were more than 200 million bushels lower than what was reported in June.
A large supply adjustment had been made in the prior marketing year as well, so the market is unclear whether this year will be three in a row. However, the trade’s recent track record is unsettling.
Sept. 1 corn stocks have fallen outside the range of estimates for the past four years, three times below and once above. Analysts severely over-guessed supply in the latest two years, boosting corn futures both times.
Sept. 1 soybean stocks have landed outside the trade range in only two of the past four years, 2019 and 2017, and the market was too high in both cases. Since 2005, Sept. 1 corn and soybean stocks came in outside of the trade range nine times each.
PAST PRODUCTION AND REVISIONS
Traders need to look out for revisions to the 2020 U.S. corn and soybean crops on Thursday. For soybeans, Sept. 1 soybean stock misses have correlated well to unexpected adjustments to the prior crop.
USDA has revised the previous year’s soybean harvest in 19 of the last 21 September stocks reports, increasing the crop 10 times and reducing it nine times. Some of those changes have been minor, but a few were impactful, especially when stocks were light as they are now.
The largest increase of the prior soybean crop in recent years was 92 million bushels for the 2007 crop, and that helped place Sept. 1, 2008, stocks 46% higher than the previous ending stocks estimate.
The largest reduction was 116 million bushels in 2019 for the 2018 crop, though that contributed to a cut of just 9% from the prior 2018-19 ending stock assumption because carryout was record-large that year.
Last year was the first where USDA’s statistics arm conducted the prior corn crop review in September along with soybeans instead of the usual December timing, which allowed the agency to review the entire 2019-20 marketing year ahead of the Sept. 1, 2020, stocks report.
Although the 2019 corn harvest was barely changed, the full supply and demand review led to the unexpected changes to the prior quarter’s stocks. That same thing happened in January 2020 when NASS reviewed 2018-19, causing an abnormally large adjustment to the Sept. 1, 2019, stocks.
The 2018 harvest in January 2020 was reduced by 80 million bushels, but the Sept. 1 stocks increased by more than 100 million bushels, and that was based on a huge cut to feed and residual use.
The trade guess of 1.155 billion bushels for Sept. 1 corn stocks would imply fourth-quarter use just below 3 billion bushels, down 4% from both last year and the three-year average.
U.S. corn exports in June, July and August were 6% above the three-year mean because of large shipments to China. Ethanol production, which accounts for more corn use than exports, was in line with the three-year average, 2020 included, after falling off pace in August.
The corn feed and residual term is the wild card because of the unmeasurable residual component, which is basically the fudge factor. June-August animal numbers were mostly below the year-ago levels, including cattle on feed and pig production.
The Sept. 1 trade prediction of 174 million bushels for Sept. 1 soybean stocks would suggest the lowest fourth-quarter use in six years at just under 600 million bushels. That would be down about 30% from both last year and the three-year average.
Atrocious June-August soybean exports landed 71% below the three-year average after a record effort in late 2020. Data from the National Oilseed Processors Association put fourth-quarter soybean crush down 5% from the three-year average.
These implied demand assumptions likely take USDA’s June 1 stocks at face value, which has proven risky in recent times.
Source: Reuters (Reporting by Karen Braun; Editing by Matthew Lewis)