Spatially varying wheat protein premiums

Many hard red winter wheat (HRWW) elevators in the northern United States test each truckload for protein and pay a premium based on the test. Other regions do not. The questions addressed here is how much of protein premiums are reflected in the local price and whether premiums are higher in areas where protein premiums are typically not paid. A spatially varying coefficient model is used to capture the variation in protein premiums implicit in prices. The data were collected by Plains Grains, Inc. (PGI), which was created to provide information to domestic and international buyers about wheat quality. They provided protein tests by elevator. Based on this information and their own tests, buyers can seek grain from elevators more likely to provide wheat with their desired protein content. We demonstrate that both wheat protein and basis are correlated across space, while the main focus is on how implicit protein premiums vary across space. A Bayesian Hierarchical Model was used to estimate a regression of basis against protein. Bayesian methods allow estimating spatially varying coefficients even with a small number of observations at each location. The focus is on how the coefficient for protein varies across space. Geospatial maps illustrate that the highest protein premiums are paid to HRWW farmers by elevators located in the western part of Oklahoma and Texas. The hedonic price of protein is significantly lower in northern states, which is consistent with protein premiums being paid directly through price premiums in these areas.