Forecasting Winter Wheat Basis with Soil Moisture Measurements

Recent literature on basis forecasting has attempted to improve upon naive forecasts by incorporating current marketing year information (in the form of basis deviation from historical levels). Given that basis is determined at each grain elevator location, inclusion of proxies for current, local supply conditions could alternatively improve upon naive forecasts. Large weather datasets have been effectively utilized in the agronomy literature to predict harvest yields in real time, however they have, to the best of our knowledge, not been incorporated into any previous basis price forecasting research. Here we match 858 weekly elevator-level, winter wheat price observations (spread over eight states and spanning 14 years) to gridded soil moisture readings (at a 0.25 x 0.25 spatial resolution) obtained from the European Space Agency and forecast winter wheat harvest basis from each weekly vantage point within the growing season. Our results indicate that inclusion of soil moisture into basis forecasts can substantially improve harvest basis forecasts for forecasts made in the middle and final part of the growing season; this result is improved upon further by disaggregating the data and forecasting at the sub-state, regional level.