An argument in favor of the development of genetically modified (GM) hybrids is that their presence is considered to be risk decreasing. On this basis, insurance premiums for corn growers in the United States who plant approved hybrids have been reduced. In this study we investigate, using a large dataset of experimental data compiled from reports of results from experimental field trials of corn hybrids by the State Agricultural Extension Services of ten universities over 20 years, whether the presence in a corn hybrid of a GM trait, or a combination of these traits, is likely to increase or decrease risk. The effects of input use on production uncertainty can be quantified through the specification and estimation of heteroskedastic production functions that allow for the variance of yield to change with the level of inputs, and we follow this approach in this study. We also use the flexible moments approach of Antle (1983) to estimate skewness of yield. We estimate a production function for the whole sample, and for three ERS regions represented in the dataset. For each model we use the residuals of the mean function to estimate the marginal effect of each input on variance and skewness of yield. The results show that there is not a systematic relationship between the presence of GM traits and variance and skewness of yield, and the results are not entirely consistent between ERS regions.