Nobel Laureate William Nordhaus demonstrates that pursuing climate mitigation will make a nation worse off


Nobel Laureate Professor William Nordhaus shows that the optimal climate mitigation policy is for far less mitigation than UNIPCCC proposes. That is to constrain warming by 2100 to 3.5°C instead of 2°C or less. But this optimal policy is based on a series of assumptions, including that policy is optimal and near universally applied. The current situation, with most countries without any effective mitigation policies, is that climate mitigation policies within a country will likely make that country worse off, even if they would be better off with the same policies were near universally applied. Countries applying costly climate mitigation policies are making their people worse off.


Last week Bjorn Lomborg tweeted a chart derived from Nordhaus paper from August 2018 in the American Economic Review.

The paper citation is

Nordhaus, William. 2018. “Projections and Uncertainties about Climate Change in an Era of Minimal Climate Policies.” American Economic Journal: Economic Policy10 (3): 333-60.

The chart shows the optimal climate mitigation policy, based upon minimization of (a) the combined projected costs of climate mitigation policy and (b) residual net costs from human-caused climate change, is much closer to the non-policy option of 4.1°C than restraining warming to 2.5°C. By the assumptions of Nordhaus’s model greater warming constraint can only be achieved through much greater policy costs. The abstract concludes

The study confirms past estimates of likely rapid climate change over the next century if major climate-change policies are not taken. It suggests that it is unlikely that nations can achieve the 2°C target of international agreements, even if ambitious policies are introduced in the near term. The required carbon price needed to achieve current targets has risen over time as policies have been delayed.

A statement whose implications are ignored

This study is based on mainstream projections of greenhouse gas emissions and the resultant warming. Prof Nordhaus is in the climate mainstream, not a climate agnostic like myself. Given this, I find the opening statement interesting. (My bold)

Climate change remains the central environmental issue of today. While the Paris Agreement on climate change of 2015 (UN 2015) has been ratified, it is limited to voluntary emissions reductions for major countries, and the United States has withdrawn and indeed is moving backward. No binding agreement for emissions reductions is currently in place following the expiration of the Kyoto Protocol in 2012. Countries have agreed on a target temperature limit of 2°C, but this is far removed from actual policies, and is probably infeasible, as will be seen below.
The reality is that most countries are on a business-as-usual (BAU) trajectory of minimal policies to reduce their emissions; they are taking noncooperative policies that are in their national interest, but far from ones which would represent a global cooperative policy.

Although there is a paper agreement to constrain emissions commensurate with 2°C of warming, most countries are doing nothing – or next to nothing – to control their emissions. The real world situation is completely different to assumptions made in the model. The implications of this are skirted over by Nordhaus, but will be explored below.

The major results at the beginning of the paper are
  • The estimate of the SCC has been revised upward by about 50 percent since the last full version in 2013.
  • The international target for climate change with a limit of 2°C appears to be infeasible with reasonably accessible technologies even with very ambitious abatement strategies.
  • A target of 2.5°C is technically feasible but would require extreme and virtually universal global policy measures in the near future.

SCC is the social cost of carbon. The conclusions about policy are not obtained from understating the projected costs of climate change. Yet the aim to limit warming to 2°C appears infeasible. By implication limiting warming beyond this – such as to 1.5°C – should not be considered by rational policy-makers. Even a target of 2.5°C requires special conditions to be fulfilled and still is less optimal than doing nothing. The conclusion from the paper without going any further is achieving the aims of the Paris Climate Agreement will make the world a worse place than doing nothing. The combined costs of policy and any residual costs of climate change will be much greater than the projected costs of climate change.

Some assumptions

This outputs of a model are achieved by making a number of assumptions. When evaluating whether the model results are applicable to real world mitigation policy consideration needs to be given to whether those assumptions hold true, and the impact on policy if violated. I have picked some of the assumptions. The ones that are a direct or near direct quote, are in italics.

  1. Mitigation policies are optimal.
  2. Mitigation policies are almost universally applied in the near future.
  3. The abatement-cost function is highly convex, reflecting the sharp diminishing returns to reducing emissions.
  4. For the DICE model it is assumed that the rate of decarbonization going forward is −1.5 percent per year.
  5. The existence of a “backstop technology,” which is a technology that produces energy services with zero greenhouse gas (GHG) emissions.
  6. Assumed that there are no “negative emissions” technologies initially, but that negative emissions are available after 2150.
  7. Assumes that damages can be reasonably well approximated by a quadratic function of temperature change.
  8. Equilibrium climate sensitivity (ECS) is a mean warming of 3.1°C for an equilibrium CO2 doubling.

This list is far from exhaustive. For instance, it does not include assumptions about the discount rate, economic growth or emissions growth. However, the case against current climate mitigation policies, or proposed policies, can be made by consideration of the first four.

Implications of assumptions being violated

I am using a deliberately strong term for the assumptions not holding.

Clearly a policy is not optimal if it does not work, or even spends money to increase emissions. More subtle is using sub-optimal policies. For instance, raising the cost of electricity is less regressive the poor are compensated. As a result the emissions reductions are less, and there cost per tonne of CO2  mitigated rises. Or nuclear power is not favoured, so is replaced by a more expensive system of wind turbines and backup energy storage. These might be trivial issues if in general policy was focussed on the optimal policy of a universal carbon tax. No country is even close. Attempts to impose carbon taxes in France and Australia have proved deeply unpopular.

Given the current state of affairs described by Nordhaus in the introduction, the most violated assumption is that mitigation policy is not universally applied. Most countries have no effective climate mitigation policies, and very few have policies in place that are likely to result in any where near the huge global emission cuts required to achieve the 2°C warming limit. (The most recent estimate from the UNEP Emissions Gap Report 2018 is that global emissions need to be 25% lower in 2030 than in 2017). Thus globally the costs of unmitigated climate change will be close to the unmitigated 3% of GDP, with globally the policy costs being a small fraction of 1% of GDP. But a country that spends 1% of GDP on policy – even if that is optimal policy – will only see a miniscule reduction in its expected climate costs. Even the USA with about one seventh of global emissions, on Nordhaus’s assumptions efficiently spending 1% of output might expect future climate costs to fall by maybe 0.1%. The policy cost to mitigation cost for a country on its own is quite different to the entire world working collectively on similar policies. Assumption four of a reduction of 1.5% in global emissions illustrates the point in a slightly different way. If the USA started cutting its emissions by an additional 1.5% a year (they are falling without policy) then it would likely mean global emissions would keep on increasing.

The third assumption is another that is sufficient on its own to undermine climate mitigation. The UK and some States in America are pursuing what would be a less than 2°C pathway if it were universally applied. That means they are committing to a highly convex policy cost curve, (often made steeper by far from optimal policies) with virtually no benefits for future generations.

Best Policies under the model assumptions

The simplest alternative to climate mitigation policies could be to have no policies at all. However, if the climate change cost functions are a true representation, and given the current Paris Agreement this is not a viable option for those less thick-skinned than President Trump, or who have a majority who believe in climate change. Economic theory can provide some insights into the strategies to be employed. For instance if the climate cost curve is a quadratic as in Nordhaus (or steeper – in Stern I believe it was at least a quartic) there are rapidly diminishing returns to mitigation policies in terms of costs mitigated. For a politician who wants to serve their the simplest strategies are to

  • Give the impression of doing something to appear virtuous
  • Incur as little cost as possible, especially those that are visible to the majority
  • Benefit special interest groups, especially those with climate activist participants
  • Get other countries to bear the real costs of mitigation.

This implies that many political leaders who want to serve the best interests of their countries need to adopt a strategy of showing they are doing one thing to appear virtuous, whilst in reality doing something quite different.

In the countries dependent of extracting and exporting fossil fuels for a large part of their national income (e.g. the Gulf States, Russia, Kazakhstan, Turkmenistan etc.) different priorities and much higher marginal policy costs for global mitigation are present. In particular, if, as part of climate policies other countries were to shut down existing fossil fuel extraction, or fail to develop new sources of supply to a significant extent then market prices would rise, to the benefit of other producers.


Using Nordhaus’s model assumptions, if the World as a whole fulfilled the Paris Climate Agreement collectively with optimal policies, then the world would be worse off than if it did nothing. That is due to most countries pursuing little or no actual climate mitigation policies. Within this context, pursuing any costly climate mitigation policies will make a country worse off than doing nothing.

Assuming political leaders have the best interests of their country at heart, and regardless of whether they regard climate change a problem, the optimal policy strategy is to impose as little costly policy as possible for maximum appearance of being virtuous, whilst doing the upmost to get other countries to pursue costly mitigation policies.


I reached the conclusion that climate mitigation will always make a nation worse off ,using neoclassical graphical analysis, in October 2013.

Kevin Marshall

aTTP falsely attacks Bjorn Lomborg’s “Impact of Current Climate Proposals” Paper

The following is a comment to be posted at Bishop Hill, responding to another attempt by blogger ….andThenThere’sPhysics to undermine the work of Bjorn Lomborg. The previous attempt was discussed here. This post includes a number of links, as well as a couple of illustrative screen captures at the foot of the table.

aTTP’s comment is

In fact, you should read Joe Romm’s post about this. He’s showing that the INDCs are likely to lead to around 3.5C which I think is relative to something like the 1860-1880 mean. This is very similar to the MIT’s 3.7, and quite a bit lower than the RCP8.5 of around 4.5C. So, yes, we all know that the INDCs are not going to do as much as some might like, but the impact is likely to be a good deal greater than that implied by Lomborg who has essentially assumed that we get to 2030 and then simply give up.

Nov 11, 2015 at 9:31 AM | …and Then There’s Physics

My Comment

aTTP at 9.31 refers to Joe Romm’s blog post of Nov 3 “Misleading U.N. Report Confuses Media On Paris Climate Talks“. Romm uses Climate Interactive’s Climate Scoreboard Tool to show the INDC submissions (if fully implemented) will result in 3.5°C as against the 4.5°C in the non-policy “No Action” Scenario. This is six times the claimed maximum impact of 0.17°C claimed in Lomberg’s new paper. Who is right? What struck me first was that Romm’s first graph, copied straight from the Climate Interactive’s seem to have a very large estimate for emissions in the “No Action” Scenario producing. Downloading the underlying data, I find the “No Action” global emissions in 2100 are 139.3 GtCO2e, compared with about 110 GtCO2e in Figure SPM5(a) of the AR5 Synthesis Report for the RCP8.5 scenario high emissions scenario. But it is the breakdown per country or region that matters.

For the USA, without action emissions are forecast to rise from 2010 to 2030 by 40%, in contrast to a rise of just 9% in the period 1990 to 2010. It is likely that emissions will fall without policy and will be no higher in 2100 than in 2010. The “no action” scenario overestimates 2030 emissions by 2-3 GtCO2e in 2030 and about 7-8 GtCO2e in 2100.

For the China the overestimation is even greater. Emissions will peak during the next decade as China fully industrializes, just as emissions peaked in most European countries in the 1970s and 1980s. Climate Interactive assumes that emissions will peak at 43 GtCO2e in 2090, whereas other estimates that the emissions peak will be around 16-17 GtCO2e before 2030.

Together, overestimations of the US and China’s “No Action” scenarios account for over half 55-60 GtCO2e 2100 emissions difference between the “No Action” and “Current INDC” scenarios. A very old IT term applies here – GIGO. If aTTP had actually checked the underlying assumptions he would realise that Romm’s rebuttal of Lomborg based on China’s emission assumptions (and repeated on his own blog) are as false as claiming that the availability of free condoms is why population peaks.

Links posted at

Kevin Marshall


Figures referred to (but not referenced) in the comment above

Figure 1: Climate Interactive’s graph, referenced by Joe Romm.

Figure 2: Reproduction of Figure SPM5(a) from Page 9 of the AR5 Synthesis Report.


Update – posted the following to ATTP’s blog


Lomborg and the Grantham Institute on the INDC submissions

Bjorn Lomborg has a new paper published in the Global Policy journal, titled: Impact of Current Climate Proposals. (hattip Bishop Hill and WUWT)

From the Abstract

This article investigates the temperature reduction impact of major climate policy proposals implemented by 2030, using the standard MAGICC climate model. Even optimistically assuming that promised emission cuts are maintained throughout the century, the impacts are generally small. ………… All climate policies by the US, China, the EU and the rest of the world, implemented from the early 2000s to 2030 and sustained through the century will likely reduce global temperature rise about 0.17°C in 2100. These impact estimates are robust to different calibrations of climate sensitivity, carbon cycling and different climate scenarios. Current climate policy promises will do little to stabilize the climate and their impact will be undetectable for many decades.

That is pretty clear. COP21 in Paris is a waste of time.

An alternative estimate is provided in a paper by Boyd, Turner and Ward (BTW) of the LSE Grantham Institute, published at the end of October.

They state

The most optimistic estimate of global emissions in 2030 resulting from the INDCs is about halfway between hypothetical ‘business as usual’ and a pathway that is consistent with the 2°C limit

The MAGICC climate model used by both Lomborg & the IPCC predicts warming of about 4.7°C under BAU, implying up to a 1.35°C difference from the INDCs, compared to the 0.17°C maximum calculated by Lomborg, 8 times the amount. Lomborg says this is contingent on no carbon leakage (exporting industry from policy to non-policy countries), whilst citing studies showing that it could offset 10-40%, or even over 100% of the emissions reduction. So the difference between sceptic Lomborg and the mighty LSE Grantham Institute is even greater than 8 times. Yet Lomborg refers extensively to the August Edition of BTW. So why the difference? There is no explicit indication in BTW of how they arrive at their halfway conclusion. nor a comparison by Lomborg.

Two other estimates are from the UNFCCC, and Climate Action Tracker. Both estimate the INDCs will constrain warming to 2.7°C, or about 2.0°C below the MAGICC BAU scenario. They both make assumptions about massive reductions in emissions post 2030 that are not in the INDCs. But at least the UNFCCC and CAT have graphs that show the projection through to 2100. Not so with BTW.

This is where the eminent brain surgeons and Nobel-Prize winning rocket scientists among the readership will need to concentrate to achieve the penetrating analytical powers of a lesser climate scientist.

From the text of BTW, the hypothetical business as usual (BAU) scenario for 2030 is 68 GtCO2e. The most optimistic scenario for emissions from the INDCs (and pessimistic for economic growth in the emerging economies) us that 2030 emissions will be 52 GtCO2e. The sophisticated climate projection models have whispered in code to the climate scientists that to be on target for the limit of 2.0°C, 2030 emissions show be not more than 36 GtCO2e. The mathematicians will be able to determine that 52 is exactly halfway between 36 and 68.

Now for the really difficult bit. I have just spent the last half hour in the shed manically cranking the handle of my patent beancounter extrapolator machine to get this result. By extrapolating this halfway result for the forecast period 2010-2030 through to 2100 my extrapolator tells me the INDCs are halfway to reaching the 2.0°C maximum warming target.

As Bob Ward will no doubt point out in his forthcoming rebuttal of Bjorn Lomborg’s paper, it is only true climate scientists who can reach such levels of analysis and understanding.

I accept no liability for any injuries caused, whether physical or psychological, by people foolishly trying to replicate this advanced result. Please leave this to the experts.

But there is a serious side to this policy advocacy. The Grantham Institute, along with others, is utterly misrepresenting the effectiveness of policy to virtually every government on the planet. Lomborg shows by rigorous means that policy is ineffective even if loads of ridiculous assumptions are made, whether on climate science forecasting, policy theory, technological solutions, government priorities, or the ability of  current governments to make policy commitments for governments for decades ahead. My prediction is that the reaction of the Grantham Institute, along with plenty of others, is a thuggish denunciation of Lomborg. What they will not consider is the rational response to wide differences of interpretation. That is to compare and contrast the arguments and the assumptions made, both explicit and implicit. 

Kevin Marshall