Daily Grain Trading: Enabled by CropProphet
Weather conditions are the dominant cause of year-to-year corn and soybean yield and production variability during the US crop season. As a result, weather conditions and events can induce grain market volatility and are a source of daily grain trading opportunities. For example, expected declines in corn production caused by the 2020 Iowa Derecho set-off corn price increases that continued for several weeks.
CropProphet users have access to a sophisticated solution to forecast end-of-season US corn and soybean yield and production, reported each year by the USDA. However, grain traders also understand that daily weather forecast updates cause volatility in US grain futures prices. CropProphet provides a “weather outlook” yield forecast that integrates the latest weather forecast information into the yield forecast. Users no longer have to guess the impact of the weather forecast on grains markets. CropProphet quantifies the weather forecast in yield terms.
Grain Futures Trading: Impact of Weather Forecasts
The CropProphet model has been reformulated for the 2021 crop season to provide improved guidance regarding the impact of weather forecast information on daily grain futures price movements. This post highlights recent analysis resulting in CropProphet 2021 product enhancements.
We’ve also examined the role of weather forecast information on grain market futures trading.
2020 Grain Futures Data Example
The analysis uses the widely available Chicago Board of Trade corn and soybean futures price information. We use a 2020 continuous contract futures price in our analysis and calculated the daily price change using the daily open price minus the daily closing price in each case.
Corn and Soybean Weather Forecast Data
CropProphet simplifies the interpretation of grain weather forecast information by integrating the weather forecast into its yield forecast model. This process converts the weather forecast into corn and soybean yield terms. A CropProphet user quickly views the impact future weather conditions may have on crop yield and production, greatly simplifying the interpretation of the weather forecast across the 1600 and 1500 counties that grow corn and soybeans in the US, respectively (see below).
Weather Outlook Grain Yield Forecast: Result
We compared the daily change of the 2020 ECMWF and GFS “weather outlook” yield forecast with the daily soybean futures prices changes between July 15 and September 15, 2020.
We evaluate the trading model with a simple evaluation.
If the yield forecast declines from the previous day, the end-of-season crop supply is forecast to decrease, and the price should rise.
If the yield forecast increases from the previous day, the end-of-season crop supply will increase, and the price should decrease.
Using this model, if the daily “weather outlook” yield forecast declines and the price increases that day or the yield forecast increases and price declines, the prediction is labeled “correct.” If both the yield forecast and price increase or both the yield forecast and price decrease, the prediction is labeled “incorrect.”
This comparison indicates the ECMWF model either correctly predicted or caused the price movement correctly 63.6% of the days. The GFS model, however, only sees 43.2% correct corn futures price movements. The green regions are the “correct” response.
The result indicates the market trades on the 00 GMT ECMWF forecast but not the 00 GMT GFS. Further investigation is underway to explore whether this is true in other recent years.
CropProphet 2021 Update
We started crop yield forecasting in 2009. Since then, we have improved the CropProphet yield forecast model each year. Some improvements come from a new data set we include in our model. Other product improvements are due to introducing new statistical techniques to the training process.
For 2021, we are focusing on improving the information available to better support daily grain futures trade opportunities. We’ve focused on ensuring the “weather outlook” yield forecast is clean and consistent compared to improvements over prior year models.