tool · draft

DICE

William Nordhaus's one-world model of economic growth and climate, used to weigh the cost of cutting emissions against climate damages and to estimate the social cost of carbon.

Weather & climate

What is reusable: partly open sourceopen datano APIclosed participation none stated; files carry a copyright notice

The Dynamic Integrated model of Climate and the Economy (DICE) is an integrated assessment model.2,3 In one model it runs from economic activity to emissions, through the carbon cycle and climate, to damages and back to policy.2,3 William Nordhaus of Yale built it; the original was published at Yale in 1992.2,4 It has been revised several times, and the latest version, DICE-2023, is described by Lint Barrage and William Nordhaus in the Proceedings of the National Academy of Sciences, March 2024.1,3 A regional version, RICE, divides the world into regions; its current version was developed by Zili Yang.1,3

The core is a model of economic growth (the Ramsey model) in which society can give up some consumption today to cut emissions and so be better off later through damages avoided.3 It is built mainly to find the best policy: it chooses the path of emissions cuts that maximises economic well-being over time, though it can also score a given policy.3 Damages grow with the square of the rise in global temperature since pre-industrial times.3 The 2023 review of studies puts them at 1.6% of world output at 3 °C of warming, though those studies leave out many impacts, such as biodiversity loss, ocean acidification and extreme events.3 For its carbon cycle the 2023 version uses DFAIR, adapted from the FAIR impulse-response model.3 Emission cuts can deepen by at most 12 percentage points in each five-year period, and every scenario assumes all countries take part with one harmonised carbon price, an assumption the paper calls highly optimistic.3

Its best-known output is the social cost of carbon: the damage, in today's money, done by one more tonne of carbon dioxide.3 DICE-2023 puts it at $66 per tonne for 2020.3 The US and other governments have used DICE to calculate this figure.2,3 The same paper projects 3.6 °C of warming by 2100 under current policies, measured from 1765.3

The model is published in a public Yale folder linked from the paper.1,3 It holds the GAMS code (DICE2023-b-4-3-10) and an Excel version kept in step with it.1 Beside them are background notes on discounting, damages, non-CO2 gases and DFAIR, and a draft third edition of the user's manual dated May 2024.1 GAMS is a commercial modelling language; the folder's notes say a licence costs around $1000, and free trials are limited in model size.1 The code is marked as copyright, with no open licence.1

DICE has no places in it, only one world economy with a single commodity.3 What it offers someone working on a place is the reasoning behind a carbon price or a social cost of carbon used to value a local project's emissions, and how much that figure moves with the discount rate and the damage estimate.3 For models that represent energy technologies and regions in detail, see GCAM, MESSAGEix and REMIND.

form
research model (GAMS and Excel)
group
Weather and climate
homepage
https://yale.box.com/s/whlqcr7gtzdm4nxnrfhvap2hlzebuvvm
cost
free to download; the GAMS version needs a paid GAMS licence

uses this engine, platform or technique

  • FaIR tool draft

    Open-source simple climate model in Python that turns scenarios of emissions, concentrations or forcing into projections of global mean temperature; fast enough to run large calibrated ensembles.

    open sourceopen dataAPIopen participation

Tags: integrated assessment, social cost of carbon, carbon price, economics, cost-benefit

staging: drafts and proposals shown