Spend Matters welcomes another guest post from Nick Peksa of Mintec.
"It has been said that something as small as the flutter of a butterfly’s wings can ultimately cause a typhoon halfway around the world.” – Author unknown
The year 1952 was when "A Sound of Thunder" by Ray Bradbury was first published. The most re-published science fiction short story of all time told the story of a hunter, Eckels. We follow Eckels as he travels back in time to a prehistoric safari and slays a Tyrannosaurus Rex, ignoring warnings of the possible catastrophic effects that changing the past could hold. Upon his return, Eckels notices a variety of changes around him and he finds a crushed butterfly on the bottom of his boot, the apparent cause of these transformations.
The butterfly effect saying is a popular one, always relating back to chaos theory. Originally a study of the behaviour of dynamic systems that are highly sensitive to initial conditions, the theory has grown to describe any situation where a seemingly small change can be magnified to produce a much stronger change in outcome.
It is possible to find these dynamic systems almost anywhere, including in the commodities market. Time and again we find a strong relationship between seemingly entirely unconnected commodities. However, often as we begin to delve deeper into the relationship between those commodities, the connection becomes apparent.
In order to see a relationship between commodities we can use what is called the Spearman correlation coefficient. This is a measure of statistical relationship between two variables. If two graphs are perfectly correlated we get a coefficient of +1. When they are exact inversions of each other we get a coefficient of -1.
Historically, there has been a very strong correlation between crude oil and edible oil prices. This relationship was in existence long before the introduction of biofuels and, therefore, is not solely due to the recent substitutability for fuel. The graph above shows the apparent relationship between crude oil and soya oil from the US. This could be shown to be a dynamic system of sorts, as a small change in crude oil prices can have a large effect on edible oil prices. This all works well, until there is excessive supply of vegetable oils or traders lose interest.
It is possible to find more unexpected correlations.
It has been suggested that oil prices were the primary cause of the recession, rather than the financial upheaval on Wall Street and around the world. Past oil price spikes associated with Middle East conflicts were each followed by a global economic recession. The price of oil doubled over the period from June 2007 to June 2008 – a far bigger increase than any of the previous spikes. It is interesting that the increase in oil price can be attributed to speculation caused by the conflict in Iraq; nevertheless these prices swiftly dropped back to levels last seen in 2007.
So, how could a rise in oil prices cause a recession? As oil prices start to elevate, so do production and transport costs, reducing financial strength and setting the table for downturns. All this shows that oil prices can affect the prices of a broad range of commodities, however seemingly unconnected. As oil prices increase, so do prices throughout the food industry.
Sometimes, however, correlations can be less than first expected.
It is easy to imagine a relationship between grain prices and meat prices, with corn being used for chicken feed, thus raising the price of chicken production. However, this correlation is not as strong as that between crude oil and edible oils, varying from 0.4 up to 0.7. Of course numerous other circumstances affect the price, such as disease, oil, and labour and transportation costs, so the correlation is reduced. The butterfly wings still have an effect here, but its force is diminished by a number of other factors.
It is important to be careful. As we’ve mentioned, it is possible to find relationships between two commodities where, in reality, none exists. You need to know your market and use a great deal of judgement to be able to interpret the correlation results correctly.