Notes on the Cost of Decarbonization

by Tom Bozzo

I took a gander at the Prometheus blog post by Roger Pielke, Jr. [*], mentioned by reader CoRev in yesterday’s guest post. Recall that the core of Pielke’s argument is that if we assume a “fixed technology baseline,” then a lot of “automatic decarbonization” doesn’t happen and the associated cost of reducing CO2 emissions skyrockets. Now, you might ask (as I did), what do the terms in quotes mean?

“Automatic decarbonization” results from autonomous processes by which economies become more energy-efficient and/or carbon-efficient per unit of output. So while the carbon emissions from world final energy use increased roughly 25% from 1990 to 2005, according to the IEA’s Worldwide Trends in Energy Use and Efficiency — from 17 to 21.5 billion tons (Gt) — the world economy expanded by roughly 66% over the same time according to the IMF. Had those emissions simply expanded with the world economy, 2005 emissions would have been some 6.7Gt (30%) greater than they actually were.

Pielke’s “fixed technology baseline” assumes the carbon intensity of the economy will not autonomously decline. In this regard, “fixed technology” is something of a misnomer; it’s more accurately termed a fixed carbon intensity baseline. [Added: Pielke notes in the comments that it’s most accurately termed “fixed energy intensity.” -TB] One argument for it is that there was a latent demand to emit that extra 6.7Gt in 2005, and the actual emissions result involved some net improvements in energy efficiency that are unavailable for additional improvements on the margin. Alternatively, it might be thought of as a variation on “past performance does not guarantee future returns.” Some of the risks along those lines are clear, for instance the Chinese and Indian economies could backslide in carbon intensity with growth of private automobile use that could swamp efficiency improvements in the US and EU.

However, as a practical matter, it’s worth keeping in mind that the recent history involved a fair amount of “automatic decarbonization” despite the bulk of the period being notable for historically cheap fossil fuels. The obvious economic implication is that there were technically feasible fuel efficiency (and by extension decarbonization) improvements that went unimplemented for being too costly at prevailing fuel prices. So the question is less whether, but how large, the autonomous reductions will be. Indeed the IEA’s energy use report totals up a few Gt of emissions reductions that could be obtained currently by propagating current “best practices” throughout the industrial and household sectors.

The fixed carbon intensity baseline raises its own nontrivial questions, namely just where the fossil fuels would come from to increase CO2 emissions by a factor of five over the next 45 years, since the scenario assumes fossil-fuel production is unit elastic with respect to output. The Peak Oil community would surely beg to differ. Saudi Arabia hopes to add a couple percent to world oil production in the next couple years. Iraq, un-CF’ed, might become more of a Saudi Arabia than, well, an Iraq. The problem is that we’d need dozens of new Saudi Arabias, and lots of new Powder River Basins, and so on, to be in a position to emit that much carbon even if for some reason we wanted to. Perhaps not surprisingly, one of the non-flood-related business stories of the week around here concerned a big order for wind-turbine parts destined for 10GW of new wind power generation in China.

There is also a conceptual problem with Pielke’s costing the entire decarbonization effort from the fixed carbon intensity baseline, at marginal costs of hundreds of dollars per ton. The autonomic emissions reductions are inherently the low-hanging fruit: stuff that pays off for other reasons and has carbon emissions reduction as side-effects costs nothing or less per avoided ton of carbon.

So Pielke’s upper figure of around 13% of world gross product really assumes that sources of CO2 emissions reduction that we know to be essentially free now and moreover are pretty big deals (e.g. efficiency improvements in the household, service, and non-household transportation sectors) actually don’t exist or are costly. You might call it an upper bound, but there are reasons to prefer least upper bounds, and this is not such an example. Is the cost of decarbonization sensitive to the amount of autonomous reduction? Sure. But Pielke’s bigger figures are on their face throwing off more heat than light.

[*] Note that Roger Pielke, Sr., is an atmospheric scientist; Pielke, Jr., is a political scientist.