Files

Action Filename Size Access Description License
Show more files...

Abstract

Past research has shown that prices move in response to WASDE reports, but have only looked at price movements right before and right after the reports. This research seeks to determine the profitability of trading based on knowing the next WASDE report at the time of the current report. The research should help traders evaluate investments in efforts to predict the report. First, a trade and hold model is used to determine the profits of trading based on whether ending stocks will be up or down in the next WASDE report. Second, a price forecast model using an ending stocks regression is used to forecast price at the next WASDE report release. The intercept of the model is calibrated so that the model predicts the current price without error; the slope is based on report data from no more than the last two years of data. Using the forecasted price, the position of the trading model’s profit calculation can change daily based on where the closing price of the commodity is in relation to the price prediction. Profits were averaged on a days-til-report, monthly, and yearly basis. Both models were profitable and the most profitable day to trade was the report release day. However, the trade and hold model outperformed the variable position model which suggests more work is needed to increase the forecasting power of this model. This might be accomplished by using additional years of data or by a form of Bayesian smoothing to improve the forecasts.

Details

Downloads Statistics

from
to
Download Full History