This looks like it could be an important paper for climate economics.
http://iopscience.iop.org/article/10.1088/1748-9326/adbd58
In short, existing climate-economics are trained with no assumption of international trade. That means that, in the models, if one country has a drought that causes food shortages it's screwed, because unlike in reality it can't just import extra food for a year or two. This means that the impact of a drought is much less than it would otherwise be, because weather is highly spatially variable.
If you retrain the models taking plausible trade networks into account, then the expected impacts from increasing temperatures is substantially worse, and since global warming is, well, global, international trade will not be as much of a solution in the future. Which means that expected GDP losses due to future warming should be much, much higher than previously modelled.
#economics #climateEconomics #climate #climateScience #IAMs #GeneralEquilibrium