ABSTRACT
Since ‘Allocation’, Nordhaus has led the way on these three attributes. This approach is in sharp contrast to important critiques of integrated assessment models, that they are black boxes, difficult to unpack and to compare (NCC, 2015; Rosen & Guenther, 2014). Nordhaus’ main assumptions are included in the paper and an appendix provides the detail. The original DICE model had 13 equations. Subsequent versions have expanded detail, but only to the extent that there are a couple of dozen now. If one disagrees with the results of the latest version, one can go to Nordhaus’ website, download the most recent code, and run it in GAMS with one’s alternative values. There is no training needed, nor any withholding of the model’s innards due to concerns about the model being abused if put in the wrong hands. On this basis, Nordhaus’ is probably the most scientific approach to integrated assessment. To be fair to the other models, credibility is likely derived from a tradeoff between transparency and sophistication, rather than simply maximizing transparency. Other models have much more detail, whether about energy technology, agriculture, the carbon cycle, and its possible feedback processes. There are so many assumptions involved that transparency becomes somewhat intractable, replication less feasible. This comes at the price of credibility. The difficulties in replicating and the immense efforts required to run insightful model comparisons raise questions about whether the detail obscures more than it enlightens or whether the detail provides different or more robust conclusions. Nordhaus appears to have a clear perspective on this question – especially when one considers that the most recent DICE model is in a much more reduced form than the Allocation model of 40 years ago.
