Sunday, December 6, 2009

A Lucid Summary of the Climategate Aftermath

Here's Mike Hammock on a lucid summary of the Climategate aftermath (HT Division of Labour's Carden, italics are mine):

I’m certainly not qualified, and ideologues won’t resolve the issue. I expect this to play out in the literature as it would in any other science. I will still defer to the collected expertise of climatologists, which, for the moment, still supports the AGW hypothesis.

I also expect climatologists to defer to economists when it comes to the question of "what to do about it". ... I also think the odds of getting a well-designed regulation out of the political system are low. The question of whether a real-world regulation would create benefits greater than costs isn't yet clear to me. I'm not an advocate of the "we must do something, this is something, therefore we must do this" position on global warming.

The problem with the climate change debate is that it has clearly become politicized to the point of acquiring millennialist tones and the status of religious dogma on both sides of the divide. It has also been captured as a rallying cry by disenfranchised socialists and communists in search of an anti-free market unifying cause. The irony here is that the current state of the politics of climate change only confirms the absolute dominance of economics as a social science: if you want to discredit free market capitalism, better to use sound economic theory and find a negative externality of gigantic proportions, and that's exactly what activists believe they've found.

For me, AGW is just another mainstream scientific theory, one that may be further validated or rejected along the track. Right now, and let's be clear about it, things are not looking good for the main AGW large-scale models. Their out-of-sample predictive power is so low that they should be considered useless for policy making (and I hope the data hasn't been meddled with to produce acceptable in-sample fit). Consider for example this peer-reviewed article (abstract below available here, italics are mine):

In their new paper in a special section of the International Journal of Forecasting on decision making and planning under low levels of predictability ... Kesten Green, Scott Armstrong, and Willie Soon asked whether it is possible to make useful forecasts for policy makers about changes in global climate up to 100 years ahead. Using evidence-based forecasting principles, they chose the "no-change" forecast as their benchmark and found that, with mean absolute errors of 0.18C for 20 years ahead and 0.24C for 50 years ahead, even if perfectly accurate forecasts of global mean temperature were possible they would not be more useful for policy makers.

They nevertheless demonstrated the use of benchmarking by assessing the relative performance of the Intergovernmental Panel on Climate Change's medium projection of +0.03C per year against historical data from 1851 to 1975. The errors from the IPCC projections were more than seven times greater than the errors from the simple benchmark that assumed no change in temperatures.

As a time-series econometrician, the argument strikes me as powerful, so I wonder: why arguments like these, based on strict scientific methodology, and published in respected journals, have been mostly ignored or downplayed in the climate change debate? The failure of mainstream large-scale models by the way is not a new phenomenon: economists once made the mistake of trusting such models, with dire economic consequences. As a result, I was taught at graduate school to always be skeptical about large-scale models. The same could easily apply to the current state of climate change research, so the question is: why is it that I don't see more skepticism and dissent in that field, particularly given evidence like the one presented above?

To make things worse, policy making in climate change has been based until now on insufficient research or poor economic understanding of the problem. It looks sometimes as if the debate is rigged to not allow for serious economic considerations or to marginalize dissenting voices - see for example what happened to Steven Levitt (undoubtedly an influential scholar) for speaking out his mind.

Things like these shouldn't be happening. By allowing activists, celebrities and politicos to act as bullies and speakers in their fields, climate researchers are steadily losing scientific credibility. I've become increasingly dismayed by the climate change scientific community as they've failed to clearly dissociate themselves from the massive use of political propaganda in their field (the movie poster on the right is a good example of propaganda that went mostly unquestioned by climate change scientists, and by the way, the movie sucks). Should I hope that the Hadley CRU scandal will serve as a wake up call for them?

No comments: