Using actual commodity prices I would like you to determine whether or not the results in Routledge, Seppi and Spatt (RSS) are reflected in the data. There are many ways to do this, most would make this homework assign intolerably difficult. So to make this manageable, I would like you just to test the qualitative implications. There are a number of them. Note that the shape of the forward curve is a function of the probability of a stock out. I would like you to test this empirically by first ranking different commodities by the probability of a stock out. You will have to figure out what the best way to do this is --- there are a number of possibilities. Perhaps the crudest measure would be the number of historical price spikes (see Deaton and Laroque). Another approach is to try to find data on inventories. You could also use the auto-correlation in spot prices, since that is related to the amount of inventory. Once you have decided on an approach, you can then test the RSS model cross-sectionally. That is, according to their model, forward curves for commodities with a different probabilities of a stock out should vary systematically across different commodities. So for example, the frequency of backwardation should depend on the chances of a stock out, etc.
I have deliberately left this open ended. You may choose the level of rigor you want to put into this. If you just want to pick a few commodities and one implication from RSS and test to see how that implication holds across the commodities by just reporting the empirical result for each commodity (and not doing a formal statistical test) that is fine. However, if you are motivated to do more, I will reward you with a higher grade.
This assignment is
due on at the beginning of class on February 15.