Richard Tol: Climate Policy Targets Revisited

  • Date: 25/04/14
  • Richard Tol, Vox

The international climate negotiations have set themselves a goal that cannot be justified.

The IPCC’s Fifth Assessment Report estimates lower costs of climate change and higher costs of abatement than the Stern Review. However, current UN negotiations focus on stabilising atmospheric concentrations of greenhouse gases at even lower levels than recommended by Stern. This column argues that, given realistic estimates of the rate at which people discount the future, the UN’s target is probably too stringent. Moreover, since real-world climate policy is far from the ideal of a uniform carbon price, the costs of emission reduction are likely to be much higher than the IPCC’s estimates.

The Stern Review of the Economics of Climate Change is the most famous economic assessment of climate policy (Stern et al. 2006). The Stern Review puts the costs of unmitigated climate change at 5–20% of GDP (now and forever), it estimates that the cost of stabilising atmospheric concentrations around 525 ppm CO2e are 1% of GDP (in 2050), and recommends that concentrations be stabilised around 500 ppm CO2e.1

The Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) finds numbers that are very different.

  • Working Group II estimates that the welfare loss due to climate change lies between 0.2% and 2.0% of GDP for 2.5˚C warming over pre-industrial times (Field and Canziani 2014).
  • Working Group III estimates that, for 2050, the costs of atmospheric stabilisation are 2.7% of GDP for a target around 500 ppm CO2e, and 1.7% around 550 ppm CO2e (Edenhofer et al. 2014).

The mandate of the IPCC does not allow it to do a cost-benefit analysis – and indeed its reports carefully avoid revealing the existence of an academic literature that does just that. Nonetheless, a qualitative comparison with the Stern Review is easily made. If the impacts of climate change are lower, and the impacts of climate policy greater, then surely optimal climate policy is relaxed. The optimal stabilisation level lies above Stern’s 500 ppm CO2e.

The political rhetoric has moved in the opposite direction. The Stern Review advocates stabilisation of the global mean surface air temperature at 2.0–2.8˚C (450–550 ppm CO2e). The international negotiations at the United Nations focus on 1.5–2.0˚C (400–450 ppm CO2e).

Figure 1 shows the results of a cost-benefit analysis, combining estimates of the Pigou tax on greenhouse gas emissions with estimates of the supply curve of emission reduction, using much the same information as the IPCC. The result is uncertain – which is why Figure 1 shows probability density functions – and depends on the discount rate – which is why Figure 1 shows three alternatives. Stern’s target is optimal only for a very low pure rate of time preference – so low in fact that the optimum is no longer a Pareto one.2 The UN targets are probably never optimal. Using a utility discount rate that is more in line with those observed in public or private investment leads to substantially less stringent climate policy.

Figure 1. The optimal concentration of greenhouse gases in the year 2100

Source: Tol (2013).

The Stern Review attracted the attention of people around the world, including Marty Weitzman, who argued that the uncertainty about climate change is such that it violates the axioms underlying expected utility maximisation – and that this justifies arbitrarily stringent climate policy (Weitzman 2009). However, Weitzman’s Dismal Theorem only holds in the absence of climate policy – if emission reduction is included in the analysis, expected utility maximisation works fine (Hwang et al. 2013, Millner 2013). Furthermore, Weitzman’s result is, strictly, not about climate policy; he finds that expected utility maximisation should not be used to analyse policy. Alternative decision criteria – minimax regret, bad tail minimisation – recommend climate policy targets similar to those suggested by expected utility maximisation (Anthoff and Tol 2014).

There is another, perhaps more serious problem with cost-benefit analysis – the optimum target is set under the assumption that it is met at the lowest possible cost. That implies a uniform carbon price – for all emissions from all sectors in all countries – that rises over time at the appropriate rate. It is highly questionable that actual climate policy will be anywhere near its cost-effective implementation. For example, there are five markets for carbon dioxide emission permits, covering part of the emissions in the EU, the US, and some developing countries. These permit markets operate on top of other regulations to reduce emissions, such as the renewables target in the EU. Elsewhere, there is a jumble of standards, subsidies, taxes, tax breaks, mandates, and so on. The actual costs of emission reduction are therefore much higher than those estimated by the IPCC.

Full comment