John Maynard Keynes, who had many impacts on study of economics and finance, has cast a long shadow in particular over the study of commodity futures, especially the relationship between the futures and the spot prices.
His 1930 book, A Treatise on Money, is remembered for its discussion of the “normal backwardation” commodity hypothesis. The underlying idea is that people and institutions that hold a lot of a physical commodity have to worry about a collapse in demand. At least in the short-run, the supply is inelastic, demand is unpredictable. As a consequence, producers have to hedge. But every producer is in the same boat, and in the ‘normal’ situation will likely estimate the dangers of demand swings in much the same way, so they won’t find many opportunities to contract with one another.
Backwardation and Contango
The upshot, then, according to Keynes, is that hedgers have to deal with speculators. Speculators can successfully demand a risk premium, in the form of a discount in the futures contract against the eventual spot price. In Keynes’ word, even if supply and demand are balanced, “the spot price must exceed the forward price by the amount which the producer is ready to sacrifice in order to ‘hedge’ himself….”
It is this backwardation, then, that is the normal situation on Keynes’ view, and the contrary situation (“contango” as it has come to be called), the one in which futures prices are higher than, and over the life of the contract decline toward, the spot price, is extraordinary.
Nonetheless, contango happens. Indeed, it happens sufficiently often to have inspired a contrary theory, the theory of storage. This view holds that inventories are sometimes very convenient, and for that reason they have value. Producers and middlemen are willing to hold inventory, and willing to incur storage and opportunity costs, in order to have what Nicholas Kaldor in 1939 was the first to call the “convenience yield” of a standing supply.
Due to this wiliness, hedgers don’t have to make the sort of desperate deal with speculators Keynes postulated, and it is contango that is the normal condition, not backwardation.
A New Look
Two Italian scholars, one at the University of Florence, the other at the University of Rome, have published a fascinating paper on the relationship between market fundamentals and commodity price volatility over time, using data from the 1920s, and data from the first decade of the 21st century, focusing especially in the case of both of these times on the prices of cotton and tin.
Their broader research project, they tell us in a footnote, involves the trading activity of John Maynard Keynes in the 1920s, and the influential views he developed on the causes and effects of volatility as a result of that experience. Keynes was especially active in the cotton and tin markets.
The Italians, Giulio Cifarelli (Florence) and Paolo Paesani (Rome) make the observation that a physical inventory of cotton or tin has a definite value, arising from the ability to meet an unexpected surge in demand. On the other hand, a physical inventory also imposes both storage and opportunity costs. According to the theory of storage, then, the difference between futures and spot prices should equal carrying costs minus convenience yield.
Their own look at the data indicates that with regard to cotton specifically, the theory of storage meets the data for both periods under review (1921-29 and 2000 – 2011).
Note that in the above graphs the lines track each other very closely. The cotton spot price, the one month to maturity futures price points, and the three month to maturity price points, as well as the corresponding figures for tin, all move up and down together faithfully, as one might expect on non-arbitrage grounds.
Upswings and Down
In accounting for such differences as there are, though, storage theory does pretty well in some periods, poorly in others. It does better on upswings than on downswings in general. During the upward price move of 1922 – 23 cotton prices behaved in the way the storage theory indicates. But during the declines in that commodity’s price both in 1926 and in the early months of 1928, the storage theory didn’t match the data, apparently due to excess stock accumulation, producing “hysteresis in the convenience yield,” Cifarelli and Paesani suggest.
In the more contemporary data, there seem to be more and longer “bouts of time” in which the storage theory doesn’t work to explain the behavior of futures prices, notably in 2001, 2003, and the whole period starting from the second quarter of 2007 and extending into 2010. Here, too, “violations of the theory of storage seem to coincide with price declines.”
In the case of tin, theory of storage does not seem to meet the data for the 1920s, though it does meet the contemporary data pretty well, perhaps they speculate due to “the highly efficient structure of the London Metal Exchange.”
Their conclusion, then, is that the markets have changed over the course of the last nine decades, and that neither a finance-driven theory, like Keynes’, nor a carrying-costs-driven theory, can be expected to explain all the phenomena.