Does Wired's Page Count Predict NASDAQ Movements?

The thing I love the most about the internet (other than the fact it provides me with a steady income!) is that I get exposed to a bunch of interesting ideas and theories that I would otherwise not hear about. One such idea making it's way through the blogosphere is the relationship between the number of advertisements in Wired, a technology and society magazine, and the NASDAQ, an American market for high tech stocks. I first learned about this relationship from the blog Marginal Revolution, though the original idea seems to come from The Podcast Network. After correcting for a mistake in his statistics, the author found the following:

I received this months Wired yesterday. I sat down for a few moments to have a flick through, and noticed two things. Firstly, it has a nice swish glossy cover, and secondly, it's fat-just like the dot.com days.

That got me thinking. I wondered how Wired's page count might reflect the technology industry. So I plotted page count against the Nasdaq, and received this surprising result.

Other than a few missing magazines, I've got most of the collection since 1996. You'll note that the Nasdaq (red) tracks Wired's page count (blue). I'm not suggesting you go an buy technology shares, but gee, I'm thinking the reports of money pumping back into technology companies might just be true given the big up-tick in this months page count (294).

It's an interesting concept and would make for a terrific econometrics project for a student looking for ideas.

Any good economic study should begin with a theory and a question with a testable hypothesis. A common theory is that the demand for labor is downward sloping - thus a rise in wages should, all else being equal, cause a reduction in the quantity demanded for labor. The question we can then test based on that theory is, "Have past rises in the minimum wage caused the unemployment rate to rise?" We can collect the data then run a number of statistical tests to see what our data tells us about the answer.

In this case our theory would be that just before or during a boom in a particular sector we should a rise in advertisements for that sector's products. Our particular question is "Does a high tech boom cause more companies to advertise in Wired Magazine?"

We've got data (though it's on the number of pages in Wired, not on the number of ads), but of course, there could be many other possible explanations why the number of ads might rise. We should try to control for these as much as possible. The obvious one, as brought up on Marginal Revolution, is that companies are probably more likely to advertise in the months before Christmas. This is something any good study should control for, by seasonally adjusting the data.

Is there anything else we might want to adjust for? If we're just concerned with fluctuations of the number of pages (or ads) and the level of the NASDAQ, we may want to de-trend the data, as there might be a natural growth trend to both. Is there other trends we might want to account for, such as the rise in personal disposable income?

There are many different directions you can take such a study in. Many economics students often think their project has to be about "classic" economic issues such as interest rates and GDP growth (I know I did when I was a student!), but if you look around the web, you can find all sorts of interesting questions (with data) to investigate. This one in particular interests me, but if it's not your cup of tea, I'm sure you can find something more to your tastes.