By Henrique Monaco and Joana Colussi et.al
Brazil’s soybean yields have grown faster than those in the United States, though this growth is not uniform across Brazil’s major producing regions (see farmdoc daily, February 20, 2025). While trend yields reflect the average rate of yield improvement over time, they do not capture the year-to-year variability that translates into production risk. This variability—measured as percent deviations from trend—offers key insights into the stability and reliability of yield growth. Building on previous analysis of Brazil’s trend yield potential and deviations (see farmdoc daily, January 12, 2017), this article examines how yield variability differs across Brazil’s main soybean-growing regions and compares these patterns to the main production regions in the U.S. Understanding relative production risk is essential for evaluating future agricultural expansion, its role in global supply chains, and the United States’s competitiveness with Brazil.
National Soybean Yields Deviates from Trend in the U.S. and Brazil
National soybean yield data from the National Agriculture Statistics Service (NASS) of the USDA and the Brazilian Institute of Geography and Statistics (IBGE) are used to assess yield variability — or yield risk — at the national levels. We first detrend each national yield series with a linear trend from 1974 to 2023 (see Figure 1). Deviations from the trend, or residuals, are then expressed as percentages of the trend yield, enabling standardized comparisons across regions with different yield levels. We then summarize yield variability using the standard deviation of these percentage deviations, providing a single metric that captures historical yield variability. This variability can be interpreted as production risk, reflecting the combined effects of weather and general growing conditions, farmers’ management decisions, and other factors that might influence yield outcomes. A higher standard deviation indicates greater yield variability and, thus, higher production risk.
At the national level, Brazil has a 10.3% standard deviation of the variability from trend. This means that in roughly two-thirds of years, actual yields fell within 10.3% of expected trend yields. For instance, if the expected yield for 2025 is 50 bushels per acre, based on historical variability, 2025 yields will likely fall between 44.85 and 55.15 bushels per acre. When comparing Brazil and the U.S. at national levels, the U.S. has lower yield variability with a 7.2% standard deviation compared to Brazil’s 10.3%. This indicates a narrower range of yield outcomes relative to trend and a lower level of production risk than Brazil. However, both are large countries with distinct production regions that face different levels of yield risk.
Regional Yield Variability Within the United States
To analyze yield variability within the U.S., we use non-irrigated county-level soybean yields from the Risk Management Agency (RMA) from 1991-2023. While not shown, we also examined yield variability using county-level yields from NASS. Differences in regional yield variability between the RMA and NASS yield data tend to occur in areas where irrigation is more widely used as the NASS county yields are not reported separately by practice. Thus, to better compare to Brazil’s regions, which are mostly rainfed, we focus on the RMA data for the comparison.
Soybean yield variability across regions of the U.S. tends to be higher than for the national average yield. In some cases the variability at the regional level is considerably higher. (see Figure 2). The low national yield variability for the U.S. is a function of the concentrated geographic production in areas with lower yield risk — such as the Midwest — and diversification effects across regions.
Source : illinois.edu