Farmers, traders, and analysts alike were left shocked after the U.S. Department of Agriculture published its optimistic corn and soybean yields last week, though the market’s narrow expectations going into Thursday’s report may carry some of the blame.
Analysts were looking for USDA to lower the U.S. corn yield to 166.2 bushels per acre (bpa) from the long-term trend of 170.7 bpa. They also predicted a decline in soybean yield to 47.5 bpa from the trend of 48 bpa.
The agency did reduce corn yield – to 169.5 bpa – above the highest of the 22 estimates collected by Reuters. However, USDA somewhat blindsided the soybean market by increasing yield to 49.4 bpa.
But should anyone have really been surprised? This is the third year in a row that USDA’s August corn and soybean yields blew out the top end of the trade range of estimates for both crops. Otherwise, that had only happened one other year in recent decades – 2001.
The ranges of analyst estimates were very narrow and specific, which is suggestive of a "group think" heading into the report. The soybean range was just 1.1 bpa, the smallest in at least 15 years. The corn guesses were the least variable in several years, and the top end of the range indicates that analysts were sure of a yield cut larger than 2 bpa.
Analysts are generally better at predicting USDA’s yields for its September update rather than the August one, but before they submit their soybean number next month, they might want to ask themselves: ‘am I too low?’
September and Final Probabilities
Historically, analysts miss the September number to a lesser degree than they do in August, but they have been missing the September soybean yield by increasing margins over the last several years, yet another hint that soybeans are being underestimated.
In the last 10 years, USDA reduced soybean yield in September four times. But all four of those reductions also followed cuts in August – clearly not applicable to 2017. And final soybean yields have not been lower than the August estimate since 2010.
Further, final bean yields over the last three years came in 5 percent, 4 percent, and 12 percent, respectively, above USDA’s trend. Considering all these factors, the hesitation of analysts to push their estimates of the final crop over trend is a bit bizarre.
There is absolutely valid reason to believe that U.S. soybean yields will not break records this year – but to not anticipate any upside from the trend seems somewhat negligent. In fairness, one of the 19 estimates provided in this category was above 48 bpa and one other matched 48 bpa, but the remaining 17 were below.
Right now, most of the U.S. soybean belt needs a good, soaking rain and weather forecasts as of late Monday suggest these rains may be on the way for many places over the next several days. If the moisture comes to fruition and the pattern stays wetter for the rest of the month, analysts might want to put a placeholder for record soybean yields in 2017 in the back of their minds.
Based on the later stage of the corn crop relative to soybeans and the current conditions, USDA may actually realize some downside to corn yield next month. The agency has reduced yield in September six out of the last 10 years.
And final corn yields have been higher than USDA’s August number only three out of the last 10 years – completely different than the tendencies for soybean yields.
The weather has been highly variable across the U.S. corn and soybean belts this summer and crop condition scores are at the lowest levels since the drought-ravaged 2012 season. This perhaps explains the mob mentality of market participants heading into last week’s report.
As of Aug. 6 – the Sunday before the August report – some 60 percent of each crop was in good or excellent shape, a few points below the normal amount. But these ratings are not alarmingly low, nor do they indicate that USDA will definitely cut yield or that the final yield will fall below trend.
The U.S. agency had twice before increased corn yield in August with crop conditions lower than the current ones, so there was some precedent for a higher bias. And the fact that weather forecasts were firm on a favorably cool bias for the first half of August should have also been a signal for analysts on both corn and soybean possibilities.
It must be considered that the root of the issue may not lie with market analysts, but rather with USDA. If the agency’s estimates have topped all market predictions three years in a row now, is it setting the initial trend expectations too low, suppressing analysts’ guesses and making the August projections look higher than they really are?
The long-term trend yields that USDA publishes are derived by the agency’s World Agricultural Outlook Board. In August, the agency’s statistics arm, National Agricultural Statistics Service, begins producing yield estimates based on both field and farm operator surveys.
Both of these groups have a tall task every year, especially now considering soybean yields have broken records in the United States for four years in a row and the last three corn yields were also the largest in history. It may be possible that these two yield-estimating procedures have stepped out of sync.
But in the nearer term, the market should avoid the mob mentality approach to estimating both USDA’s moves and the final numbers, because it will reduce the shock effect to the market and add more value back into the polls.
The opinions expressed here are those of the author, a market analyst for Reuters.