We didn’t produce this much only to produce this much

A framework for understanding yield in agriculture

Apr 19, 2023

Hello and welcome the fifth edition of Topsoil – a monthly newsletter with frameworks to help you make sense of agriculture, at just the right depth.

Before we dig into yield today, I wanted to share a couple podcast appearances over the past month:

I joined The Modern Acre for a second time to dive deeper on the Business of Farming. If you haven’t tuned into Tim and Tyler’s podcast, I highly recommend it as I always learn something new from their other guests. I’m already looking forward to chatting yield with Tim and Tyler next month!
Rachel and Ciara from the Luminexus podcast and I chatted about agriculture, and why it is the coolest industry for STEM students to be considering. We had a blast talking about both the status quo today and what the future of ag may hold.

Thank you for your continued support in reading, subscribing and sharing!

Last time, we talked farm-level profitability. However, we barely dipped our baby toe into the income side of the equation: “Revenue from crop products is a function of yield multiplied by price.” Simple, right?

Today, we are going to unpack the metric that often rules everything around it: yield. We’ll cover what it is, what drives it, how far we’ve come, and what our ever-higher-yielding future might look like.

Yield is a measure of how much crop is produced across an area of land. For example, corn yield in the US is measured in bushels per acre, as in “this is 200 bushel corn” where you say the “per acre” part silently in casual conversation.

Depending on where you are in the world, or what crop you are talking about, the units may vary. Instead of bushels, some countries may measure in kilograms or tons, and hectares instead of acres. Some crops are more commonly measured in pounds or by unit.

So if yield is how productive the crop is, what drives yield?

Unlike a factory, where inputs are mechanically churned into a reliable number of widgets, crop yield is a function of countless, interdependent factors. Agriculture takes place in a dynamic, living, breathing system.

The Genotype-by-Environment-by-Management framework (GEM for short) can help make sense of these factors and how they interact to influence yield.

Genotype x Environment x Management (GEM) is a framework to understand how the drivers of yield interact.

G is for Genotype

You, me, each individual corn plant, and every other living organism on Earth has a unique sequence of DNA, called a genotype. The genotype has a big impact on yield because it provides the “blueprint” for the plant – it determines all of the features that an individual plant can possibly have. Once the plant is growing in a real environment, those features that you actually observe in the field are called the plant’s phenotype.

Using a non-plant example, both Great Danes and teacup poodles are dogs. The genotype of a Great Dane means it has a high likelihood of weighing well over 130 pounds (59kg). Teacup poodles, on the other hand, possess a genotype that sets the potential size likely under 10 pounds (5kg). The phenotype of the Great Dane or teacup poodle is how much each individual dog actually weighs.

Similarly, every major crop like corn, rice, or apples has vast potential genetic diversity within each species. They can be tall or short; have more or fewer leaves; larger or smaller fruit – you get the idea.

It is up to plant breeders to develop new varieties with features, or traits, that we as humans find appealing. Things like: resistance to disease, sweeter flavors, bigger kernels, faster ripening, and on and on.

Looking at the last 100 years of improvements to yield (which we’ll dig into deeper below), genotype is responsible for a whopping 70% of the increases.

While seed breeding is critical, it is wildly expensive. It often takes over 10 years and $5M to develop a single new seed variety for any of the major commodity row crops.

And, like a teacup poodle that ends up weighing 15 pounds instead of the expected 10 because of a treat-rich lifestyle, the genotype “blueprint” interacts with the environment and management to form the what-we-actually-see phenotype.

E is for Environment

The environment is simple in concept – it is the crop’s surroundings, like soil, water, sunlight, air quality, temperature and weather. Plants need many of these ingredients in just the right amounts to produce a harvestable crop.

However, “just the right amount” doesn’t always happen. Sometimes there’s too much water or too little, it’s too hot or too cold – all of which can limit yield. We call these yield-limiting factors “stressors”. Biologists lump environmental stressors into two buckets:

Biotic stressors: These are living things that can harm yield, such as weeds, non-beneficial insects, fungal or bacterial diseases, and other pathogens.
Abiotic stressors: These are non-living things that can harm yield, such as flooding, temperature extremes, salinity, or drought.

Of course, the environment varies based on where you are in the world. The combination of soil, rain and sunlight in Iowa, US or Mato Grosso, Brazil contributes to a much higher soybean yield than, for example, the Sahara Desert.

With climate change, the environment has been shifting over time, in many cases creating conditions that are less favorable for crop production. While there is some research that indicates we haven’t seen total yield drop due to climate change yet, climate change is expected to reduce yield, with certain regions being hit particularly hard.

Finally, like Miley Cyrus circa 2010, the environment can’t be tamed. Farmers can only do their best to manage the environment, which brings us to our last piece of the GEM framework.

M is for Management

So what’s a farmer to do?

In the GEM framework, Management refers to all of the choices and actions that a farmer takes to influence yield.

This includes: selecting the right seeds or genetics, controlling biotic stressors like weeds, pests and disease, mitigating harms from abiotic stressors by fertilizing and irrigating, and making hundreds of other decisions on crop rotation, planting and harvesting.

As we covered with the crop spectrum, each crop requires slightly different management techniques. For example, in almond orchards, management decisions (like irrigation set-up) made in the first year impact yield for years or even decades.

To make things more complicated, when it comes to management, 1+1 does not always equal 2 and it is often unclear how much each decision or practice contributes to the total yield. Each decision interacts with every other decision or practice, making it tough to untangle cause and effect.

In recent years, there is optimism that we can better tease apart this messy web. Advances in machine learning and more powerful statistical modeling are helping researchers understand and predict how each decision contributes to yield. As for now in 2023, we are still in the early days – there is no crystal ball or all-knowing AI-bot yet.

A historical view on yield

We love a chart that goes up and to the right, especially when it showcases progress for us as a human species. 📈

Take a look at crop yields for a few major crops over time:

Per acre yield over time for US corn, wheat and soybean (1866-2021). Source: USDA Crop Production Historical Track Records April 2022.

From the 73 years between the end of the US Civil War in 1866 when the USDA began keeping record, to the beginning of WWII in 1939, corn yield remained flat.

With that track record, imagine telling someone in 1940 who had experienced yield hovering at the same level for 3 generations, “over the next 80 years, corn yield will increase by over 500%.”

You would be right and no one would believe you.

Looking closely at the chart above, there have been two big watershed breakthroughs that have catapulted yield to the highs that we enjoy today.

First, in the late 1930’s, US farmers began to rapidly adopt hybrid corn. Instead of the open-pollinated varieties that were previously the status quo, hybrid corn was an intentional cross (and sometimes double-cross) of two different varieties of corn. Hybrid corn plants were hardier and produced more grain – resulting in about 0.8 bushels/acre more each year from this improvement in genetics.

Next, in the mid-1950’s, along with the continued improvement in genetics, farmers began to adopt other management techniques that further drove up corn yield by 1.9 bushels/acre/year. WWII ushered in broadscale production of nitrogen fertilizer, originally ammonium nitrate for ammunition. Additionally, pesticides and mechanization for the war effort were re-routed to farming after the war ended.

The yield increase over time is not a phenomenon unique to commodity row crops. When we look at a specialty crop like apples over the past 100 years, we see a similar gentle up-and-to-the-right curve of progress:

Per acre yield over time for Washington, US apples (1919-2021). Washington grows over half of the apples in the US. Source: USDA NASS.

While corn, wheat, soybean, and apple yields have all shown improvement, corn takes the cake for most improved. This got me wondering, what makes corn so special?

Surprisingly, it doesn’t seem to be the sheer number of acres covered, as corn and soybeans have roughly the same acreage (at least in the US).

Instead, recall that genotype is responsible for 70% of the yield improvements over the past 100 years. Most of the corn exceptionalism can be explained by simple biology and R&D budgets.

A friend (thanks Dan!) pointed out something obvious once you see it, but subtle if you are not in the business of plant breeding. Apples, soybeans, and wheat have male and female reproductive parts right next to each other – each flower or floret contains anthers (male) and an ovary (female).

If you are a plant breeder trying to cross two different varieties, you have to carefully (we’re talking literal tweezers here, folks) make one plant “female” by removing every single anther of each flower, so only the ovary remains. Then, you have to carefully dust pollen from your “male” plant on each flower that had its anthers removed. Voilà! That flower will produce seeds for a new hybrid or cross.

Corn is different. Each corn plant still has both female and male parts, but those parts are in completely separate structures. The pollen-producing tassels are at the top of the plant while the female part (the ear) is further down the stalk. This biological happenstance has made a world of difference for plant breeders over the past 80 years. They can simply remove the top part of the corn plant to make it “female” for breeding purposes. Imagine how many more crosses you can make when you don’t have to get out the tweezers every time!

Perhaps as a result of the biology above, there is simply more R&D for corn than for other crops. A lot more. When looking at applications to the USDA for field testing (a proxy for R&D efforts) over a 15 year period, corn accounted for 43% of the applications. The next closest crop was soybeans with only 8% of the applications.

Mind the yield gap

Tom Brady. Serena Williams. Tiger Woods.

You could say that David Hula occupies a similar echelon when it comes to corn yield. He holds the record for growing the highest yielding corn, clocking in at 616 bushels/acre.

To hit such astonishingly high yields, David Hula does everything he can within the GEM framework on his farm in Virginia. He selects high yielding seed genetics, masterfully pulls all the management levers that are available, and happens to farm in an environment with abundant sunlight.

The year that David Hula hit the 616 bushel/acre record, the US average was 171 bushels/acre. While this 445 bushel difference can be attributed to many factors and not every environment or genotype can even dream of producing 616 bushel corn, there is a very similar concept that drives a large swath of innovation in agriculture: the yield gap.

Researchers measure the yield gap by first calculating the yield potential of each different environment. Then, they observe the actual yield harvested in each of those environments. The yield gap is equal to potential yield minus actual yield.

While more of an academic exercise instead of something that farmers do in their day-to-day operation, the yield gap represents opportunities for improvements in genotype or management. With each improvement, the yield gap decreases.

This closing-of-the-gap is a critical unlock for the future of agriculture. By farming more efficiently on farmland that already exists, we can support a growing and wealthier population that often demands more animal protein. Importantly, we can do so without converting natural environments to farmland (like by razing rainforests or plowing up prairies).

To illustrate this point, the graphic below shows the decreasing amount of land it takes to produce 10 bushels of corn over time:

Source: Bayer’s 2022 Sustainability Report

As with everything in agriculture, closing the yield gap is complex. Shane Thomas from

Upstream Ag Insights

had an extremely thoughtful take on this point and the graphic above. Shane walks through the evolution of key management practices, such as the reduction of crop protection used per acre, while acknowledging lingering challenges like fertilizer runoff in waterways.

Yield for the future

I just talked for 2,000 words about how the abundant rivers of grain coming off fields is a wonder of human achievement. Critical for us to feed billions around the globe while protecting the environment. That this is the paradigm driving agricultural progress for nearly a century.

And yet, what if we are thinking about yield in a very one-dimensional way?

Today, we think of yield as volume from one crop in one growing season. This is what farmers are paid for and what makes the agricultural world go round.

However, new technology, consumer demands, and policy may shape how we think about yield in the future:

Not just volume: Instead of being paid based purely on volume, it is possible that players throughout the value chain could pay for other features of a crop instead, such as nutritional density or for specific macro- or micronutrient content.
Not just one crop: Precision agriculture is rapidly advancing, making it plausible that farmers of the future will be able to manage their crops at a plant-level at scale (instead of a field or zone level today). With that shift, might we see yield measurements also advance to take into account intercropping or even integrated crop-livestock systems on each acre?
Not just outputs: As the prices of inputs have increased, many farmers already carefully consider not only the output of yield, but the inputs used and total profitability. Along these same lines, perhaps in the future, other measures such as nitrogen use efficiency or carbon intensity are more heavily weighted.
Not just production: While the sheer volume produced is a critical piece of the equation to improve food security around the world, it is certainly not the whole picture. The distribution of what is produced matters a lot – both in terms of where grain is moved around the world, as well as what it is used for (directly feeding humans, feeding livestock, fueling vehicles as ethanol, etc.). This may be more of a stretch, but perhaps the way we measure yield in the future takes into account the end use.

As the saying goes, “we didn’t come this far to only come this far.” We didn’t produce this much only to produce this much. I’d love to hear your thoughts on other ways yield might evolve in the future!