I am looking for feasible model(s) that reflect dependence of a forward price of a agricultural commodity on weather conditions. The goal is to model how a change in the forward price reflects a change in current weather conditions. For simplicity, one can assume supply/demand conditions are considered separately, i.e. do not change. Any references? Thanks.
You just need to use statistical analysis… that is, run a regression of the commodity price on an intelligently selected set of independent variables.
That’s a tough one because as you stated supply and demand will matter, as will acres planted. Additionally it will be the cumulative effect of weather conditions across several regions (Brazil, Argentina, Asia, Africa, Russia and North America) that will have to be I guess given a weighted average. I guess you’d want to pull cumulative precipitation levels while also putting a factor in accounting for how lumpy the distribution for the rainfall was and accounting for average temps. Many regions have multiple planting seasons so you will have to run these separately for each harvest, weight that by expected commodity type (probably based on acres planted estimates) and account for late or early harvest starts which will be typically a factor of temps and / or excessive rainfall that can postpone planting, although the start date could be an input assumption. I’m sure the models are out there, but your average layman typically just looks at USDA crop and farm surveys and estimates which cover most regions and are pretty up to date and comprehensive.