Support the Arctic Sea Ice Forum and Blog

Author Topic: Damage Functions In Integrated Assessment Models  (Read 2852 times)

rboyd

  • Nilas ice
  • Posts: 1334
    • View Profile
  • Liked: 226
  • Likes Given: 52
Damage Functions In Integrated Assessment Models
« on: June 22, 2017, 09:46:41 PM »
I was reading a paper by William Nordhaus which laid out some of the assumptions used in his DICE IAM. What struck me was the incredible simplicity of the Damage Function (which calculates the economic losses due to climate change) and the vast understatement of impacts inherent in it.

The Damage Function used is the square of the temperature, in degrees, multiplied by 0.236% of GDP. With this, 3 degrees produces a 2.1% reduction in global GDP and 6 degrees produces an 8.5% loss. Given that the general scientific consensus is that 6 degrees would be the end of civilization as we know it, doesn't seem to me that 8.5% of GDP quite captures the impact correctly.

Given that these IAMs are being used for policy decisions, for example the calculation of the cost of carbon, such spectacular shortcomings are severely worrying. I have linked to the Nordhaus paper below, as well as a research paper that looks at some of the shortcomings of IAMs. Would be useful if there are any individuals on the forum who are involved in the construction of IAMs.

Some of the discussions remind me of the financial modelling prior to the 2008 crash, simply extrapolating from past experiences, non-stochastic distributions, and assuming that "everything else remains the same" such as fully functioning and deeply liquid markets. The calculation of losses in IAMs very much share these issues. Of course, there is also the underestimation of geophysical impacts in the GCMs but that is a different topic.

http://cowles.yale.edu/sites/default/files/files/pub/d20/d2057.pdf

http://www.impactlab.org/wp-content/uploads/2016/04/Empirically-calibrating-damage-functions-and-considering-stochasticity-when-integrated-assessment-models-are-used-as-decision-tools.pdf



rboyd

  • Nilas ice
  • Posts: 1334
    • View Profile
  • Liked: 226
  • Likes Given: 52
Re: Damage Functions In Integrated Assessment Models
« Reply #1 on: June 24, 2017, 07:11:21 PM »
Good paper on the shortcomings of IAMs and the way in which the damage function is constructed

"One of the most important parts of an IAM is the damage function, i.e., the relationship
between an increase in temperature and GDP (or the growth rate of GDP). When
assessing climate sensitivity, we can at least draw on the underlying physical science and
argue coherently about the relevant probability distributions. But when it comes to the
damage function, we know virtually nothing – there is no theory and no data that we can
draw from. As a result, developers of IAMs simply make up arbitrary functional forms
and corresponding parameter values."

"IAMs can tell us nothing about the likelihood or possible impact of a catastrophic climate
outcome, e.g., a temperature increase above 5°C that has a very large impact on GDP.
And yet it is the possibility of a climate catastrophe that is (or should be) the main driving
force behind a stringent abatement policy."

http://web.mit.edu/rpindyck/www/Papers/PindyckClimateModels2015.pdf

rboyd

  • Nilas ice
  • Posts: 1334
    • View Profile
  • Liked: 226
  • Likes Given: 52
Re: Damage Functions In Integrated Assessment Models
« Reply #2 on: June 25, 2017, 07:53:27 AM »
Bocconi Centre for Research on Energy and Environmental Economics and Policy: Estimation of climate change damage functions for 140 regions in the GTAP9 database

Utilizes climate model forecasts (somewhat outdated given the more recent scientific papers) to assess impacts from sea level, and increased temperatures. Doe show the great disparity between individual country outcomes at 3 degrees, but still a relatively simple assessment. Does not take into account second order effects, nor issues of overall societal stability/instability, nor extreme values (e.g. short periods of heat extremes that may kill / severely damage plants as against average temperature increases).

ftp://ftp.repec.org/opt/ReDIF/RePEc/bcu/papers/iefewp86.pdf
« Last Edit: June 25, 2017, 07:59:56 AM by rboyd »

rboyd

  • Nilas ice
  • Posts: 1334
    • View Profile
  • Liked: 226
  • Likes Given: 52
Re: Damage Functions In Integrated Assessment Models
« Reply #3 on: June 25, 2017, 08:19:21 AM »
2017 National Academies of Sciences: Valuing Climate Damages: Updating Estimation of the Social Cost of Carbon Dioxide

Specific Chapter on the Damages Modules of the three models: DICE, FUND and PAGE. All of the report can be downloaded for free. This diagram shows the low level of losses estimated for increases in temperature in 2100 which many would equate to societal collapse (the damage amounts are compared to an assumed 2100 GDP of $370 trillion):



https://www.nap.edu/read/24651/chapter/8


« Last Edit: June 25, 2017, 08:24:30 AM by rboyd »

rboyd

  • Nilas ice
  • Posts: 1334
    • View Profile
  • Liked: 226
  • Likes Given: 52
Re: Damage Functions In Integrated Assessment Models
« Reply #4 on: June 29, 2017, 09:25:38 PM »
The Social Cost Of Carbon

Great background on the use of IAMs to calculate the social cost of carbon, and the huge issues with such models. As Carbon Brief says, probably better just to pick the required emissions level and set a cost of carbon that gets a society.

The forcing of economic cost benefit analysis upon all regulatory initiatives in the US during the Reagan era represented a great win for the opponents of regulation. Given the exclusion of so many "non economic" costs and benefits, high discount rates etc. it tilts the playing field against regulation.

https://www.carbonbrief.org/qa-social-cost-carbon