By GianCarlo Moschini, Yongjie Ji, and Seungki Lee
Production agriculture depends heavily on exogenous environmental conditions, including weather. As such, agriculture is acutely vulnerable to the deleterious long-run effects of climate change. Indeed, mounting evidence suggests the likelihood of large negative impacts. What can be done about it? Actions to deal with climate change can be thought of as pursuing “mitigation” and/or “adaptation”—mitigation is about containing climate change by reducing greenhouse gas emissions, whereas adaptation blunts and counteracts the damaging consequences of climate change. Countries’ free-riding incentives make global cooperation to reduce emissions difficult, and thus mitigation problematic. Adaptation, by contrast, is less vulnerable to opportunistic behavior because investments in adaptation often have local payoffs and substantial private good aspects.
Technologies to foster agricultural climate change adaptation may include new crop varieties with traits that enhance resistance to pest, disease, and environmental stress (e.g., heat tolerance and drought resistance). Varieties with shorter growing cycles and earlier maturation, precision agriculture technologies, and more efficient water management and expanded irrigation are also expected to be helpful. Ultimately, all this requires major R&D investments, from both the public and private sectors, to support enhanced innovation efforts in adaptation-enabling new technologies.
What is the scope of such an R&D challenge? To shed some light on this question, we have studied the impact of weather and technology on US corn yields. Corn is the most important field crop in the United States, and one that has benefited greatly from major technological advances over the last few decades, including the development and widespread adoption of genetically engineered (GE) varieties. The latter constitutes the most prominent set of agricultural innovations since the green revolution, and were made possible by large R&D investments by agrochemical and seed companies. GE traits have been shown to have cost-reducing and yield-increasing effects. Because the nature and scale of the “GE revolution” in agriculture are well understood, we propose using the yield impact of GE traits as a “yardstick” to gauge the scope of the innovation task required for adaptation to climate change.
By combining various data, we first estimate the impact of weather variables and technology on US corn yields from 1981 to 2016. The estimated yield model, along with weather projections from mainstream global climate change models, permits us to forecast expected yields at mid-century and end-of-century, and thus characterize the implied “yield gaps” due to anticipated climate change. We can then compare such yield gaps with the size of the one-time yield gains due to first-generation GE traits in corn production.
Figure 1 illustrates results for one mainstream climate model (HadGEM2-ES) and one warming scenario (RCP4.5)—we actually have results from twenty climate models and two warming scenarios, which paint a more nuanced picture, but this figure will suffice to highlight the main ideas. Average yield realizations from 1981 to 2016 show considerable variability due to changes in year-to-year weather conditions. Yet, there is a clear upward trend in this productivity metric, reflecting the impact of continuing improvements in technology. Specifically, our model teases out the separate impact due to (first-generation) GE traits from the underlying productivity gains attributable to all other sources. We find that, by 2016, adoption of GE traits is associated with yield gains of about 16 bushels/acre (compare points A and B in the figure). We also find that the functionality of GE traits matters—specifically, it is the insect-resistance traits that are responsible for observed maize yield gains, while there appears to be no evidence of yield benefit from the adoption of herbicide tolerance traits. The residual yield gain from the underlying technical progress, separate from the adoption of GE traits, is estimated at about one bushel per acre per year (average across all US counties). Furthermore, accounting for weather conditions in the model is essential in order to identify the role of technology in maize production. We find that yields are significantly positively impacted by growing degree days, are negatively impacted by excess heat, and are sensitive to precipitation and water stress.
The estimated model is then used for counterfactual simulations to determine the expected yield impacts of anticipated climate change at mid-century (2040–2059) and end-of-century (2080–2099). This step relies explicitly on the future weather conditions projected by climate models. In figure 1, the dashed blue line illustrates expected future corn yields assuming the continuation of the underlying technical change and no climate change—climate variables here are set at the predicted averages of the historical 1980–2005 period. The red line, by contrast, illustrates expected future yields conditional on projected climate change. Note, however, that we are still assuming continuation of the underlying technical change (approximately one bushel per year, other things equal). The impact of first-generation GE traits is considered fully captured by 2016—had GE varieties not been developed and adopted, the yield forecast would follow the bottom, dashed gray line.
It is apparent that the model predicts sizeable yield shortfalls due to the projected weather changes. Figure 1 illustrates that the “yield gap” at mid-century, in this scenario, is about 21 bushels/acre (distance between A’ and C’), and this gap rises to about 61 bushels/acre by the end-of-century (distance between A” and C”). Comparing such yield gaps with the model-estimated yield gains due to GE varieties provides a useful characterization of the innovation-adaptation challenge posed by climate change. As figure 1 illustrates, the yield gap due to climate change at mid-century is about 2.6 times the entire yield gains made possible by the adoption of GE varieties. By the end of the century, this yield gap is almost four times as large as the GE yield bump.
The GE productivity gains that our model captures relate to a clearly defined set of innovations—so-called first generation GE varieties embedding agronomic traits—that were rapidly diffused (essentially to full adoption) over a relatively short and well-defined time period. Beyond exemplifying a success story of agricultural innovation, this provides us with a useful yardstick. We know what it took for the GE revolution—it was made possible by propitious scientific breakthroughs in basic science and molecular genetics at public institutions, and supported by massive R&D investments by industry. Similarly, successful adaptation of agriculture to climate change is likely to require purposeful, directed investments in R&D to develop suitable new technologies. The scale of this innovation challenge is not trivial. Our study suggests that large and sustained research efforts might be required to counter the negative implications of anticipated climate change in agriculture.
For more information, see:
Lee, S., Y. Ji, and G. Moschini. 2021. “Agricultural Innovation and Adaptation to Climate Change: Insights from Genetically Engineered Maize.” CARD Working Paper 21-WP 616, Center for Agricultural and Rural Development, Iowa State University.
Suggested citation:
Moschini, G., Y. Ji, and S. Lee. 2021. "Corn Yields and Climate Change: The Innovation Challenge." Agricultural Policy Review, Winter 2021. Center for Agricultural and Rural Development, Iowa State University. Available at www.card.iastate.edu/ag_policy_review/article/?a=116.