Battle Rejoined on Climate Causing African Wars

♠ Posted by Emmanuel in ,, at 9/09/2010 01:27:00 AM
While the publication Nature usually features stuff that interests research scientists rather than social scientists, this news item may be of interest to IPE followers as well. Two years ago, there was a well-publicized book called Climate Wars: The Fight for Survival as the World Overheats that made exactly the case that climate change triggered conflicts. In line with this idea, the economist Marshall Burke and his colleagues at UC Berkeley empirically investigated the relationship between indicators of climate change and conflict. Here is the abstract of their downloadable paper from last year:
Armed conflict within nations has had disastrous humanitarian consequences throughout much of the world. Here we undertake the first comprehensive examination of whether global climate change will exacerbate armed conflict in sub-Saharan Africa. We find strong historical linkages between civil war and temperature on the continent, with warmer years leading to significant increases in the likelihood of war. When combined with climate model projections of future temperature trends, this historical response to temperature suggests a roughly 60% increase in armed conflict incidence by 2030, or an additional 390,000 battle deaths if future wars are as deadly as recent wars. Our results suggest an urgent need to reform African governments' and foreign aid donors' policies to deal with rising temperatures.
So far, the pop notion of climate change triggering conflict is given some empirical support. However, a newer paper by Norwegian political scientist Halvarg Buhaug of the Peace Research Institute finds otherwise. Here's the abstract from his likewise downloadable paper in which he says climate is not to blame for Africa's civil wars:
Vocal actors within policy and practice contend that environmental variability and shocks, such as drought and prolonged heat waves, drive civil wars in Africa. Recently, a widely publicized scientific article appears to substantiate this claim. This paper investigates the empirical foundation for the claimed relationship in detail. Using a host of different model specifications and alternative measures of drought, heat, and civil war, the paper concludes that climate variability is a poor predictor of armed conflict. Instead, African civil wars can be explained by generic structural and contextual conditions: prevalent ethno-political exclusion, poor national economy, and the collapse of the Cold War system.
Which is correct? Beats me, pal. I haven't had the time to look over the methods and whatnot to form an opinion like, say, the one which made dubious claims that IMF structural adjustment caused tuberculosis deaths in Eastern Europe. As with many of these econometric studies, there are definitional issues: On the independent variable side, what constitutes an incident of climate change? On the independent variable side, what constitutes and incident of civil war? In its writeup Nature starts off with the later study...
But is there real proof of a link between climate change and civil war — particularly in crisis-ridden parts of Africa — as many have claimed?

No, says Halvard Buhaug, a political scientist with the Peace Research Institute Oslo in Norway. In research published today in Proceedings of the National Academy of Sciences1, he finds virtually no correlation between climate-change indicators such as temperature and rainfall variability and the frequency of civil wars over the past 50 years in sub-Saharan Africa — arguably the part of the world that is socially and environmentally most vulnerable to climate change. "The primary causes of civil war are political, not environmental," says Buhaug.

The analysis challenges a study published last year that claimed to have found a causal connection between climate warming and civil violence in Africa. Marshall Burke, an economist at the University of California, Berkeley, and colleagues, reported a strong historical relationship between temperature and the incidence of civil war. They found that the likelihood of armed conflict across the continent rose by around 50% in unusually warm years during 1981-20022. Projected future warming threatens to offset the positive effects of democratization and eradicating poverty in Africa, they warned.
...before discussing quarrels on method mentioned above:
The two rival groups are now disputing the validity of each other's findings.

Buhaug says that Burke's study may have been skewed by the choice of climate data sets, and by their narrow definition of 'civil war' as any year that saw more than 1,000 fatalities from intra-national conflict. The definition is at odds with conventional measures of civil war in the academic literature, says Buhaug: "If a conflict lasts for 10 years, but in only 3 of them the death toll exceeds 1,000, [Burke et al] may code it as three different wars...You'd really like to apply as many complementary definitions as possible before proclaiming a robust correlation with climate change," Buhaug adds.

Burke maintains that his findings are robust, and counters that Buhaug has cherry-picked his data sets to support his hypothesis. "Although we have enjoyed discussing it with him, we definitely do not agree with Halvard on this," says Burke. "There are legitimate disagreements about which data to use, [but] basically we think he's made some serious econometric mistakes that undermine his results. He does not do a credible job of controlling for other things beyond climate that might be going on."

Buhaug disagrees vigorously. "If they accuse me of highlighting data sets in favour of my hypothesis, then this applies tenfold more to their own paper."
I think regression models are too blunt a tool for answering this question. Disaggregating political and social variables was already hard enough before they threw ecological ones into the picture. Interaction effects, anyone? Judgement will inevitably play a role as this issue will become an increasingly more important one in national, regional, and international policy circles. Stata may help, but it isn't a substitute for clear-headed thinking.