According to USDA, the average U.S. soybean yield was 35.3 bushels/acre in 1995, the year before GM herbicide tolerant varieties were made commercially available. By 2016, with 94 percent of U.S. soybean acreage planted to GM herbicide tolerant varieties, average soybean yield had risen to 52.1 bushels/acre. However, crop yields vary from year to year and over time due to many factors. Year to year variability is largely due to differences in weather, while trends may be due to increased yield potential of available varieties through traditional breeding, changes in management practices, and the availability of inputs, larger scale climate change, and other factors. Looking at multi-year averages can smooth out year to year variability. In the 10 years before GM soybeans became available, the average U.S. soybean yield was 34 bushels/acre. From 1996-2005, years when adoption increased to 87 percent of soybean acreage, the average U.S. soybean yield was 38.69, and from 2006-2016, the average yield was 44 bushels per acre. These increases are unlikely due solely to the introduction of GM herbicide tolerant soybean varieties, as U.S. soybean growers had effective weed management options prior 1996. The benefits of GM herbicide tolerant soybean varieties to U.S. soybean growers have been primarily the simplicity and flexibility of the program as well as some cost savings.
The most recent data available from USDA directly comparing GM and non-GM soybean yields in the U.S. is from 2006, when GM herbicide tolerant soybean adopters were found to have a three percent yield advantage over non-adopters on average. Again, however, it is important to remember that average differences may be driven by other factors than the GM herbicide tolerance trait, such as other characteristics of the varieties being grown and/or management ability of farmers, etc., which may be systematically different between adopters and non-adopters. These types of systematic differences between adopters and non-adopters may lead to either over- or under-estimating the impact of GM herbicide tolerant alone when looking simply at average differences.