I feel that I must first admit that I haven’t actually finished reading Porter’s The Competitive Advantage of Nations. I have it on my shelf right now but I haven’t gone past chapter two mainly because of my other academic obligations. Not to mention that it’s a very thick book. Probably thicker than War and Peace! (I haven’t finished reading that one either)
Anyway, an interesting thing I learned about cluster theory is that it is actually a theory based largely on correlation rather than hard statistical facts that indicate causation. While the theory was built out of an empirical study of numerous industries from ten nations, I have been told that these were really case studies that didn’t delve deeply enough into the data.
The thing with correlation is that it does not imply causation. To base a theory on mere correlation would make it potentially weak and unable to stand up to hard tests. As we’ve seen in my previous post, there are some problems with the theory and it needs a hard re-think. To add to that, I also got an early preview of Stephen Chen’s work relating to cluster theory. The finding of that study was that cluster theory, in its current form, does not stand up at all to statistical tests. Basically, it’s premature to say that clustering firms together causes an improvement in their performance. He even goes so far as saying that the relationship might be the other way around or even spiral.
I would probably rate this as the most interesting personal discovery of the week. Considering that cluster theory is being used by many governments these days, you would think that they thoroughly scrutinised it first.