In a previous post we have tried to debunk long only commodity investing. The main arguments why it does not work over long time frames is because of the curve structure associated with commodities that have to be stored in warehouses or silos. Below we show the annualised return of long only positions consisting entirely of the commodity shown as a function of percentage of time spent in backwardation. We distinguish between those commodities that are part of the GSCI and those that are not. We superimpose a linear fit to all the data points. Notice the positive slope associated with the fit. This shows that an increased time spent in backwardation leads to greater annualised returns. Below we only consider data from 1 Jan 1990 onward. Returns are calculated on roll adjusted price series and transaction costs are not included.
The GSCI is a production weighted index designed to reflect the relative significance of each of the constituent commodities to the world economy. These types of commodity indices are notoriously highly skewed in favour of energies. The weightings for the different sectors for the calendar year 2020 are shown in the plot below. The plot above and below shows us that the index will only perform well when the commodities in the energy sector are in sustained periods of backwardation.
In this section we reconstruct the GSCI by taking the current weights as a proxy for the historical weights. The value of $1 invested in the spread in represented in the plot below.