The UnAustralian

Tuesday, May 20, 2003
Misleading and Irrelevant

The title of this post comes from my description of Castles and Henderson's critique of selected IPCC emission scenarios (IPCC SRES). The title is also a partial tribute to Sylvain Galineau from Chicagoboyz who describes my description as "name calling and bullying". Something which gives me a great deal of amusement given that he called the IPCC's emission scenario report "gigantic pile of academic elephant shit". Given that Sylvain's and my debate is currently going around in circles, I don't really feel like carrying on this treadmill, so instead I'll defend my use of the words "misleading" and "irrelevant" as accurate descriptors of the Castles/Henderson critique.

* Misleading

In my opinion, Castles and Henderson present a very misleading picture of the IPCC emission scenarios. If you are familiar with the scenario's themselves, then it isn't misleading, however they chose to initially present their arguments in the public domain, rather than in scientifically peer reviewed literature (apparently it will be published soon in Energy & Environment). In a nutshell, the IPCC presents a number of scenarios with high and low estimates of economic growth. Castles/Henderson focus on two high growth scenarios (A1 and B1), at the exclusion of the low growth scenarios. I feel that their claim is misleading as the B1 scenario is consistently pointed out as being the scenario which has the lowest emissions. The reason why this is misleading is that far too many people confuse low emissions with low gdp growth. Nothing could be further from the truth.

While gdp does play a role in the total level of emissions, there is also another (more important) factor; technology. As the world reduces the proportion of energy produced from carbon dioxide emitting sources (a process which has been occurring for the past two centuries), emissions will drop. To use a hypothetical scenario, if complete decarbonisation could be achieved, it wouldn't matter how much energy we used. Now, given that high rates of growth, and technological change tend to go hand and hand, it's not particularly surprising that (at least some) of the high growth scenario's go hand in hand with low emissions.

A similar example of being quite deceptive is this quote for Castles "This is a far faster rate of growth in output per head than has ever been achieved by the developing world in the past, even for a single decade. Some economists may believe that living standards in developing countries COULD increase by a factor of 9 from 1990 levels by 2040 - but none of them, to my knowledge, believes that this marks the lower bound of the range of future possibilities. And many scientists believe that average living standards in developing countries CANNOT increase by a factor of 9 - by 2040 or ever (see, for example, this year's Presidential Address to the American Association for the Advancement of Science by Dr. Peter Raven)." The problem with this is that the scenario which Castles was referring to, B1 IMAGE, isn't the lower bound. Far from it. The range of growth rates was 2.3 and 5.2%. The B1 IMAGE scenario has a rate of growth of 4.5%. Much much closer to the higher growths than the lower growths. (Thanks to Sam Levikman for finding this quote).

My complaint with Castles/Henderson is that the way which they present their case is misleading, and has led many intelligent people to make incorrect statements. For example, on the popular site Kuro5hin, a story submitter wrote (after mentioning many tales of very high growth:

"According to these two professors, using the IPCC's most conservative estimate of GDP growth rate of developing countries, all of the above instances will happen."

Or for an example closer to home, Sylvain has written:

"And that, because GDP is assumed to drive GHG emissions, has a significant impact on the final result over such a long-term projection."

Once again, wrong.

Given that the test for "misleading" should be determining whether or not it does mislead intelligent people, I think that it can be pretty conclusively said that it is misleading.

* Irrelevant

My second claim is that the Castles/Henderson critique is irrelevant. It really makes no difference to the real world whatsoever.

Castles/Henderson have attacked the IPCC for using gdp($US) rather than gdp($PPP). If the IPCC was comparing standards of living, then they would have a good point. But they aren't. Rather there is a clear historical relationship between gdp($US) and energy usage. This relationship has applied to all regions of the world (except for the Middle East and North Africa*) over the last 19 years, and in the US (where more long term data exists) 190 years. The IPCC have assumed that this relationship will hold (essentially less energy per dollar will be required, as gdp increases, as the economy will develop towards being more service based, and more efficient technologies will be introduced).

Essential it can be written that energy usage is a function of gdp($US). Likewise, energy usage is probably a (different) function of gdp($PPP). Hence the most logical indicator to chose is the one with the most supporting data. And guess what, it's gdp($US). One of the lead authors of the IPCC SRES has pointed this out to Castles and Henderson, but they appear to have taken the John Lott approach (don't acknowledge the criticism).

Likewise, another problem that Castles/Henderson have brought up, is that certain countries have massive gdp growth rates in the IPCC scenarios. This is simply an interesting effect of modeling regions rather than individual countries (the effect occurs when a rich country is in the same region as poor countries, so rich country gets the benefit of the higher growth rate of the poor countries). This is an interesting statistical anomaly (which only occurs if you assume that growth rates are uniform within a region - an assumption which isn't needed by the IPCC SRES). However, despite this effect, there is nothing wrong with the regional data** (it isn't compiled by adding together the various growth rates of individual countries, but rather by treating each region as a giant super-state). For those who are interested in the economic models, this site lists details of the IMAGE model (one of the models used by the IPCC).

* Interesting PPP are good at comparing locally produced goods (food etc), but not so good for comparing goods purchased on the international market (cough... oil... cough).

** A comparison between the regional rates of growth and historical rates of growth can be found here.
| 11:39 PM