Tuesday, 3 June 2008

Economists, oil, cricket and correlation

One of my bug bears, and something that I am increasingly aware of editing a maths magazine, is the bad use of maths to prove a point, particularly regarding cause and effect.

Yesterday, I was reading The Independent on Sunday, not a newspaper usually prone to sensationalist journalism and easily my favourite UK newspaper (there is not all that much competition, it doesn't have breasts on page 3).

Hamish McRae writes on economics, is a very well respected economist and a fine writer, and writes with some insight. He was writing on the current rise in oil prices, and presented a graph of the oil price between 2000 and 2008. On this graph, the Nasdaq technology index between 1992 and 2000 was also plotted. It showed a startling similarity between the two, and at first glance you would think that oil prices are due for a crash, just like the dot-com burst of 2000.

Now, I don't believe that McRae was trying to be misleading at all - he was simply making the point that markets often overshoot in the way that they price things. However, thanks to the way the page was edited, a casual observer would read this page and predict an oil price crash.

Of course, the problem is that not only does the graph present different time frames and so different market conditions, but oil and technology stocks are completely unrelated. You could never use the Nasdaq to predict oil prices, a comparison is completely spurious. We all need energy, but we don't all need technology shares. They make look similar, but there is no cause and effect and the factors underlying the pricing are completely different.

So, I thought to myself, what else could we plot against oil price. Let's check out the Australian Cricket Team's Test winning percentage averaged over its last 40 games between 1982 and 2004.

It's all cricket's fault

As you can see, the correlation is remarkable! So, when you fill up this week and notice the latest price rise, you can blame Allan Border, Mark Taylor and Steve Waugh for improving the results of the Aussie Test Team! By the way, the data are completely real - make it for yourself at Cricinfo.

Economists are prone to this kind of thing - not the insightful ones, like Hamish McRae, but those without mathematical understanding. Books about how economics explains everything in our lives are a dime a dozen these days, from the very good ones like Freakonomics, to others I have had the misfortune of reviewing. They are prone to interesting ideas, like how the greedy free market will solve global warming, to how economic growth is good for all, both the poor and the rich. Did you hear the one about the two economists who paid each other $10000 to eat a poo sandwich? The GDP of each economist went up by $10000. Yes, economic growth is always a good thing.... Buy me a beer, I'll go on for days with my half-baked economic theories!

I will write a more thorough story on this over at Plus soon.