The UnAustralian

Friday, May 30, 2003
And Again...

The last point of Sylvain's post is probably the most interesting, and worth quoting:

Predicting one single country's GDP 100 years ahead is in and of itself too ludicrous to contemplate. Even given 21st century tools and computing power, a 1900 French economist trying to predict his nation's 2000 GDP would have most likely been buried by his own implicit and explicit assumptions. Most of the population then, for instance, lived from the land in the countryside. It is quite unlikely our scientist would have correctly assumed 80% of his countrymen would live in cities at the end of the century, and even assuming he did, what it meant for GDP figures. If only because most of 2000 France's economic activity, and the jobs that go with it, did not exist back then. My job did not exist when my parents were born. The tools I use did not exist when I was born. Today's children will be able to choose careers that didn't exist when they were born. Try and model that. You can't, of course. So you have to assume away all the unknwon of a whole century. And that's your margin of error. And like interest rates, it compounds over time to staggering figures that make the entire exercise pointless. Ken would probably laugh at anyone who would try and predict the level of the Nasdaq in 2100. But somehow, a bunch of 2100 global temperature averages are credible even though the hopelessness and relevance of the exercise are identical.

While the hypothetical French economist would be the luckiest man in the world had he correctly predicted France's economic future 100 years from now, he could have taking a range of realistic growth projections and obtained a range of possible futures. Trends such as urbanisation (or decarbonisation - to use a more relevant example) have been going on for centuries. In all likelihood, they will continue.

I don't have data on France's economy now and then, but it is illustrative to look at another country, the US. To take a 50 year period (1948 - 1998), the per capita GDP grew at around 3.5% (I'm deliberately skewing this numbers by not including WWII - the inclusion of this high growth period would support the SRES projected growth rates even better, whereas going back further would lower growth rates because of the depression). This would put it about right in the range of growth figures used by the IPCC.

The important point here, is that while it is next to impossible to predict the future accurately, it is considerable more easier to construct high and low end cases. And in all likelihood, reality will fall somewhere in-between them.
| 1:42 AM