Wednesday, June 01, 2005

Predicting climate change

I've been playing the Foresight Exchange game for a few months. It's an internet-based "Ideas market" in which players try to increase their "wealth" (sadly, it's only virtual money) by betting against each other on future events. This idea was pioneered by Robin Hanson and I've talked about how I believe it could be a useful addition to the climate change debate elsewhere.

Most of the market is politics and technology based, but there are a few environmental issues covered too. Here I'm going to take a quick look at one of them - a prediction of the atmospheric CO2 level in December 2030 (ok, this is really a prediction about human behaviour, technology and politics more than the environment per se, but it has direct environmental relevance).

The "claim" pays $0.01 for each ppm by which the atmospheric CO2 level exceeds 400ppm in December 2030 (with a ceiling of $1 for 500ppm or more), so the current mid-market price of $0.675 implies a prediction of 467.5ppm. Note that inflation and discount rates play no part in this market. I've plotted the Mauna Loa CO2 below along with a few projections.



The thin green line is an exponential fit to the last 30 years of Mauna Loa data, extrapolated out to 2030. The annual increase in CO2 (currently a little less than 2ppm/year) grows by about 1.4% each year, compounding up to about 438ppm by 2030. These values do change a bit depending on how many years of data are fitted, but still always fall short of the predicted value.

The red and cyan lines are projections based on a generous 2ppm/year in 2000, increasing by 2% or 3% per annum thereafter (the data for the 4 years 2000-2004 shows an actual increase of just below 8ppm). As you can see, even the 3% curve falls short of the market's predicted 467.5ppm. Note that these x% projections are NOT simply an x% rise in the atmospheric CO2 level, they are the percentage increase in the size of the annual increase (which is currently running at around 2ppm = 0.5% of the total). So even if this annual increase grows at a rate of 3%, more than double the historical level, the prediction will not be reached.

My conclusion is that the claim is significantly overpriced, and that something in the region of 450ppm (=$0.50) or even lower would be much more realistic. Disclaimer: I hold a short position in this claim.

1 comment:

Anonymous said...

I created that claim and I always thought it was trading high.

I considered creating a claim about emissions rates that could be arbitraged against CO2LVL. The price of 67 implied very high emissions or some sort of runaway effect like CO2 releases from melting tundras.