Corn fills wagon

After five back-to-back years of greater-than-normal-trend corn and soybean yields, 2019 seems positioned to experience less-than-normal-trend conditions. Less-than-expected yields coming after a cold wet spring also created significant prevented-planting conditions. But the magnitude of production hiccups -- yields and acreage -- are unknown. The 2019 corn-yield guide is a starting point for benchmarking 2019 corn-yield projections.

Consider historic U.S. corn yields

Figure 1 is a graph of U.S. average corn yields from 1960 to 2018. Yields during that time frame are shown by the trend line. They have increased at an average annual rate of 1.87 bushels per acres. That increasing trend in corn yields throughout time is thanks to a whole host of technological improvements. But the trend makes it challenging to compare reported yields from year to year, even during a short time frame. In just five years the increasing trend in yields is equal to 9.35 bushels -- 5 years x 1.87 bushels.

The data from Figure 1 is also shown in Table 1. The table format makes it easier for those interesting in looking at the specifics. The first thing many will notice is the five-year run of greater-than-average-trend yields in recent years. That's something only seen previously in the late 1960s. Using that data method -- detrending yield since 1960 -- readers would find that the trend average yield for 2019 is 171.6 bushels per acre.

The U.S. Department of Agriculture has made two corn-yield estimates for 2019. The first projection was 176.4 bushels per acre and assumed normal spring-planting conditions, which 2019 didn't have. In the more recent World Agricultural Supply and Demand Estimates reports for June and July, the USDA used a lesser-yield projection of 166 bushels per acres.

Table 2 attempts to provide historical context to current yields projections. Specifically the departure from trend data in Table 1 has been ranked from least to greatest. Then each year’s departure from trend has been added to the previously discussed trend average yield of 171.6. That creates a 2019 yield benchmark.

In the middle of Table 2 is a dark black line. That denotes the 50th percentile – or where half the observations are greater or less than average. Note that greater-than-average-trend yields occur more than half of the time -- about 61 percent to be precise. If readers search the column on the right for 166, they will be near 1964 and the departure from the trend of negative-5.8 bushels per acre. Historically speaking a final 2019 yield of 166 would be the 14th-least departure from trend in 60 years, of the 23rd percentile.

The goal here is not to provide a yield forecast, but to help provide context to any yield forecasts that might be taken out of consideration. Using yield data since 1960, the current USDA projection of 166 is yield at the 23rd percentile. As one thinks about yield less than that number, it's worth considering how historically significant such a yield might be. Again none of that is to say how plausible any yield forecast is. It's merely to help provide context to how such yields fit into historical context.

Lesser yields may be increasing

Every so often we hear a version of the argument that lesser yields are not possible anymore given improvements in technology. The idea is that genetic improvements have made significant strides, especially on the bottom end. Therefore a year like 2012 wouldn't have 2012-like yield impacts. More specifically if 2012 yields are detrended to 134 in 2019, the argument is that 2012 weather would have yields greater than 134 given the significant improvements during the past few years.

During the cold wet late planting of 2019, we heard “the bad isn’t as bad as it once was” argument surface again. That line of thinking, which was even included in some agriculture-media stories, caused us to wonder what the data might suggest.

To think about that we considered the data in Table 2 and created three categories based on yield departures from trend – the worst 25 percent yields, the middle 50 percent and the best 25 percent. Once each year was categorized into a group, a trend line was calculated from posted yields -- USDA reported yields -- for the best and worst yield groups.

A trend line drawn through the best 25 percent of yields, in gray, shows an average increase in yields of 1.87 bushels per year. That's compared to the trend line for the worst 25 percent of yields, in blue, of 1.61 bushels per acre per year.

Admittedly one could argue that methodology is a bit simple. That said, those data and methods don't support the ideas that the worst yields are catching up faster and limiting the downside yield risk in trend-adjusted terms. While that topic area deserves more research, one might conclude from Figure 2 that the worst corn yields are increasing at a slower rate throughout time. In turn the downside risk for yields in any given yield could be understated.

Wrap it up

As we approach the time of year where yield estimates and forecasts are thrown around a lot, it’s important to keep in mind the context of those yields. First of all, yields since 1960 have increased at an average annual rate of 1.87 bushels per year. That makes it necessary to detrend yields to make historical comparisons.

When thinking about the USDA’s current 2019 yield forecast of 166 bushels per acre, it’s worth considering that the yield is at the 23rd percentile of historical yields. Again that's not to say anything about the quality of the forecast, but to note and benchmark where those current yields rank in a historical context.

Finally we heard many comments this spring about how technological improvements have made big strides to increase lesser-end yields. When comparing the trend lines for the worst 25 percent to the best 25 percent, it’s difficult to argue the worst yields are catching up. One could even make the argument that they have been decreasing, and down-side risk has been underappreciated.

Visit for more information.

David Widmar is an agricultural economist with Agricultural Economic Insights.