Kate Mackenzie Climate change costs: Too much is not enough

Is avoiding climate change just too expensive?

Answering this question depends very on how much you think climate change itself will cost, and how much you think avoiding climate change will cost -  along with difficult moral and philosophical matters such as concern for future humanity.

One of the best known attempts to quantify this is the Stern Review, which estimated that climate change could end up costing 5 – 20 per cent of global GDP  per year, but would only cost 1 per cent of GDP to avoid (Lord Stern later revised these numbers, saying 2 per cent of GDP is needed to avoid climate change, and without this, the effects could be even worse than he previously thought).

Enter Bjorn Lomborg. Best known as a climate change sceptic, he now says an agreement on climate change at Copenhagen in December is crucial. But he’s not entirely going with the mainstream just yet: he believes the dominant approaches to climate change are too expensive, so the latest project of his Copenhagen Consensus Center is to publish a series of papers on the best ways to spend $2,500bn over the next 10 years, and then ask five leading economists to choose the best option.

The latest of these papers is likely to be one of the most controversial. Richard Tol, an energy economist at Dublin’s Economic and Social Research Institute and a member of the Intergovernmental Panel on Climate Change, estimates that the costs of most climate-change mitigation programmes vastly outweigh the benefits the world would gain.

Tol came up with five scenarios for how to spend that money (some scenarios spent more or less):


Tol came up with five scenarios for spending $2,500bn over 10 years:

1. A $750/tonne carbon tax introduced only in the OECD, over 2010 – 2019. Cost per dollar of benefit: >$100. Carbon concentration: about 875ppm

2. A $250/tonne carbon tax inroduced worldwide, over 2010 – 2019. Cost per dollar: $100 Carbon concentration: about 750ppm

3. As for number 2, but continuing the carbon tax throughout the century (thereby spending far more than the $250bn). Cost per dollar: $50. CO2: 450ppm

4. A $12/tonne carbon tax, the proceeds of which are invested in a trust fund for 10 years, to finance a century-long programme of emission abatement. Cost per dollar: $4. CO2: 650ppm

5. A $2/tonne carbon tax, the proceeds of which are invested in a trust fund as above. This spends only $12.5bn.  Cost per dollar: 66c. CO2: 850ppm.

One important note about the scenarios: apart from scenario 1, they all assume a uniform, worldwide tax on carbon across all sectors. There’s really almost no chance of this happening any time in the foreseeable future, let alone from next year – for anything like this to play out in reality, those costs would have to be concentrated in some countries.

Tol’s report says “The fifth scenario is the only project worth funding.” But when pressed about the resulting 850ppm atmospheric carbon concentration, he was more ambivalent:

“That is the level that is justified by a conservative cost-benefit analysis. You can make the argument that this analysis ignores uncertainties and inequities. So if you correct for those things, perhaps you go for a carbon tax that is five to 10 times as high – $12, then we go for 650ppm, which may be much more palatable.”

How can 850ppm even be considered remotely palatable, when scientists believe it would wreak havoc on the environment? The IPCC’s best estimate puts this effect at 2.4 degrees celsius in 2100, but Tol’s literature review finds the cost of climate change between 2 and 3 degrees is only equivalent to a few per cent of income in 2100 (in fact he says there are actually net benefits for parts of the world between 1 and 2 degrees). There are two important factors here: Tol’s models deflate the future cost of climate change using a high discount rate, and ignore uncertainty, which is a key reason scientists warn a change of more than 2 degrees is unfeasibly dangerous:

Discount rates

The discount rate is one of the most contentious points here, and was the aspect of the Stern Review that was debated most heavily. A discount rate addresses the difference in costs applied in the future. Spending money in the future isn’t as costly as spending money now, partly because as a society, we expect to be wealthier in the future, and partly because of our own impatience: we prefer not to have to spend the money now, so we value it more in the present than we do in the future. Tol uses a discount rate of 5 per cent, which he points out is used by HM Treasury and by markets (interest and inflation are also commonly a component of a discount rate).

Plenty has been written about the discount rate used by Stern, but suffice to say that Tol’s paper values the cost of climate change impacts in the future at far lower than does Stern.

Uncertainty and missing impacts

Apart from the discount rate, the other big contentious point here is that it does not account for the uncertainty of climate change impacts. Tol actually spends several pages on the subject of  “missing impacts” not quantified in the papers on cost he reviews. These range from saltwater intrustion onto groundwater to reduced labour productivity – though he argues that many of these are small, and that positive impacts are also “missed”. He concludes (our emphasis):

The “missing impacts” are a reason for concern and further emphasize that climate change may spring nasty surprises. This justifies greenhouse gas emission reduction beyond that recommended by a cost-benefit analysis under quantified risk. The size of the “uncertainty premium” is a political decision. However, one should keep in mind that there is a history of exaggeration in the study of climate change impacts.

Over the course of writing this story, I spoke and corresponded with Dimitri Zenghelis, who contributed to the Stern Review. The lack of ‘uncertainty’ was one of his biggest objections to Tol’s methodology:

The reason scientists and others are calling for early action on emissions reductions is to insure against the risk of catastrophic, irreversible, climate events like widespread floods, droughts, storms, heatwaves, famine, disease and devastating coastal inundation, which would render billions of people poorer than they are today. Using outdated science and loaded assumptions, Tol’s study instead assumes that climate change can only ever have a marginal impact on GDP affecting richer generations in the future. This is bad economics, divorced from any scientific basis.

The case for early action

But what about all those cost-effective – even money-saving – efficiencies that we keep hearing about? Tol is sceptical about this:

Because of decreasing returns to scale, doubling the rate of technological progress means that the effort that is being put into improving energy efficiency has to be more than doubled. This is easy to do for a specific technology, but hard across the entire economy. Furthermore, only a fraction of appliances, vehicles and machines are replaced each year. That is, technological progress applies to a fraction of the capital stock only. Premature retirement of capital is very expensive.

However Zenghelis says this overestimates the cost of efficiencies:

In fact, up to a third of the necessary reductions can be achieved without any carbon price at all, through better regulations and standards on efficiency in buildings, domestic appliances, vehicles and production processes. Over a twenty-year period, this will have a profound effect on the capital stock working with the replacement cycle.

Tol also argues that renewable energy has only established niche applications during times of high energy prices, and has “never captured the market”. Zenghalis argues that forcing earlier action will lead to faster innovation, as new technologies are scaled up they either fail or undergo sharp cost reductions. This needs supportive policy, he admits, but “Technological innovation and policy action induce each other, hence the cost-effective positive feedback that makes the path chosen so important.”

But the big reason for not delaying, Zenghelis says, is simply that we can’t afford to wait:

With every year of delay, the stock of GHGs rises. To stabilise the stock, emissions must fall to a level equal to the earth’s natural capacity to absorb the gases (inflows in to the system must equal outflows, think of GHG stock as a water tank where emissions flow in). Consequently, the pathway for emissions has to be lower (not just relatively to a baseline but in absolute terms) in order to stabilise at a given temperature for each year that action is delayed.

Best policy

At the end of his report, Tol returns to the ‘uncertainty’ factor:

Given the large uncertainties and the large inequities of climate change, one may justifiably
argue that the “right” policy is more stringent than the “optimal” policy shown here.

However, he says the fact that his scenarios assume that spending would be perfectly allocated – which would certainly not be the case in real life – support the opposite case:

However, the analysis presented here also omits suboptimal policy design. Carbon price differentiation and direct regulation may well increase abatement costs by a substantial margin. Therefore, one should perhaps not go too far beyond the optimal policy outlined here.

In otherwords, spending would in reality be far less efficient, and yield a far worse cost-benefit, than his models suggest.

To which Zenghelis says:

He is quite right to note that poorly applied, uncoordinated, emissions reduction policies can significantly raise mitigation costs, perhaps even doubling them. But it is highly misleading to imply that omitting this in some way offsets the gaping chasm of an omission left by ignoring climate uncertainty. Poor policy implementation costs are of second order when compared to the uncertainties of climate change.

Sorting it all out

Much of the disagreement over these ideas comes down to the tricky technical, moral and philosophical questions of the costs of climate change decades into the future, and whether the risk of not doing enough is bigger than the risk of wasting money. As John Quiggan of University of Queensland writes in an unpublished paper on the debates over Stern’s discount rate:

“The real difficulty here is that we are pushing economic analysis to its
limits… Economists can help to define the issues, but it is unlikely that
economics can provide a final answer.”

But the Copenhagen Consensus papers are not a million miles away from other economists on some aspects of this series. One of Lomborg’s main points, when I spoke to him about this paper, was that more cost-effective approaches to avoiding climate change are not being given adequate attention in all the focus over CO2 emissions and cap-and-trade schemes. Three of these – methane, forestry/afforestation, and soot, were discussed in papers accompanying Tol’s report. All three fared better on this cost-benefit analysis than CO2 abatement through efficiencies and alternative energy.

*Footnote: carbon tax versus CO2 tax

My newspaper story refers to a CO2 tax rate, which is what is usually discussed in Europe under the emissions trading scheme there, but the Tol study itself talks about a carbon tax.

Carbon: $250 – CO2: $68

Carbon: $12 – CO2: $3

Carbon: $2 – CO2: $0.50

Related links:

Soot, methane, and forests: Cost-benefit analysis (FT Energy Source, 14/08/09)
Climate change report urges cut in costs
(FT, 14/08/09)