Controversies in Climate Change Economics
by
Robert Eastwood
Department of Economics
University of Sussex
July 2010
Abstract
This paper is a non-technical review of the Economics of global policy on reducing greenhouse gas emissions. Quite a lot is known about the likely physical consequences of anthropogenic climate change, but much uncertainty remains. In particular, account needs to be taken of possible catastrophies such as ice sheet melting. How are we to balance the known costs in the present of taking action to reduce greenhouse gas emissions against the uncertain benefits of such action to future generations? How convincing is the case for substantial action now? If the case for such action is accepted, should emissions be controlled via Kyoto-style national emissions targets or by the imposition of carbon taxes? How can the challenges of burden-sharing between developed and developing countries be addressed?
Keywords
cap-and-trade, carbon tax, climate change, emissions permit, greenhouse gas, Kyoto Protocol
Biography
Robert Eastwood is Senior Lecturer in Economics at Sussex University, specializing in Development Economics, in particular the links from demography to economic growth and poverty in developing countries. He was an expert witness to the 2006 All Party Parliamentary Group on Population Development and Reproductive Health and is a member of the Steering Group of the current African Economic Research Consortium project on Health, Growth and Poverty Reduction in Africa
Prepared for the first issue of Environment and Society.
Word count excl. refs and title page 8837. Five figures.
Introduction
The judgements in the IPCC’s fourth annual report (FAR) regarding anthropogenic global warming can be summarized as follows:
(a) The warming trend in the last half of the twentieth century amounted to 0.13 °C per decade, or 0.65 °C for the fifty-year period, and it is ‘very likely’ that most of this increase was due to anthropogenic GHG emissions. If atmospheric GHG concentrations stayed at the 2000 level, estimated warming in the 21st century would be a further 0.6 °C. [i] [IPPC 2007a, pp 30, 39 and Table 3.1, p.45]
(b) In the absence of policies to reduce emissions (i.e. under ‘business-as-usual’), estimated warming in the 21st century is between 1.8 °C and 4.0 °C, according to which of six scenarios is chosen, but each of these estimates is subject to substantial uncertainty, so that the overall range of warmings that are judged ‘likely’ in one scenario or another (see fn.1) is between 1.1 °C and 6.4 °C. [ibid. Table 3.1]
The science behind these judgements involves three main steps. First, the climate-relevant impacts of human action (principally GHG emissions) must be estimated. Second the so-called radiative forcings (RF) generated by these impacts must be estimated. This terminology arises from the fact that each impact can be expressed numerically in terms of the change in the intensity of incoming solar radiation that would have the same effect – for instance, because of the greenhouse effect, the impact of doubling atmospheric CO2 is equal to 3.8 Watts per square metre, equivalent to a rise of about 1.7% in incoming solar radiation. [ibid. Table 2.4, Shaviv 2006] Third, climate sensitivity, defined as the effect on equilibrium global temperature of a given RF must be estimated. Since the RF of doubling atmospheric CO2 is known, climate sensitivity can be neatly re-expressed as the rise in equilibrium global temperature that such a doubling (from the pre-industrial level of about 280 ppm) would produce. The FAR gives the ‘likely’ range for climate sensitivity so defined to be 2 °C to 4.5 °C and considers it ‘very unlikely’ that it is below 1.5 °C. [ibid. p.38]
For the Economist seeking to analyse climate change policy, the key lesson from the above is that the amount of anthropogenic warming under business-as-usual is almost certainly positive, but it is also subject to a great deal of uncertainty. This uncertainty is both economic and scientific. Economic uncertainty is reflected in the six scenarios considered in the FAR, which vary according to what is assumed about economic growth, population growth and the speed and direction of technological advance. [ibid. p.44] Corresponding to these scenarios are atmospheric concentrations of GHGs equivalent to CO2 concentrations in 2100 ranging from 600 ppm to 1550 ppm [ibid. Table 3, note (c)]. Scientific uncertainty, indicated by the wide range of likely values for climate sensitivity, arises from various sources, of which by far the most important is uncertainty about the link between warming and the formation of clouds of different types and at different altitudes. [ibid. Table 2.4]
One further source of uncertainty should be mentioned. Any case for action on climate change depends not on the anthropogenic component of warming discussed above, but on total warming. If anthropogenic warming happened to be against a background of natural cooling at a comparable rate, then - far from being a threat - it would be viewed as delivering us from a new Ice Age.[ii] That such a notion may not be entirely fanciful follows from the established fact that for the past million years, the earth has experienced a series of Ice Ages lasting around 100,000 years each, interspersed with interglacial periods, warmer by some 5 °C, lasting around 10,000-15,000 years. [Kunzig and Broecker 2009, IPPC 2007b, King 2006]. It is from this perspective that one feature of current warming becomes vital – its rapidity. As noted above, the upper end of the ‘likely’ range for anthropogenic warming in the 21st century is 4.5 °C, corresponding to about 5 °C over the 150 year period starting in 1950. Although our understanding of natural temperature changes over such short time scales is imperfect, we have no scientific basis for expecting any significant compensation from natural processes over the next one or two centuries, so a rough equation of warming with anthropogenic warming is reasonable.
Given the scientific background, arriving at an assessment of climate policy requires a number of steps. But the case in principle for global policy action is straightforward. It is that GHG emission creates a global externality. A private action which releases a gram of CO2 into the atmosphere imposes future costs on other persons across the globe which are not taken account of by the emitter, and which are the same wherever and however the emission takes place. Externalities mean, in general, that uncoordinated private actions produce bad outcomes.
To illustrate this principle in the simplest possible way, consider the case of a single emitter and a single victim. The emitter will rationally increase its level of emission until the gain derived from the last unit of emission falls to zero. The victim of emission, however, will (usually) be willing to pay something – say $1 - to avoid that last unit, so a negotiated agreement whereby the victim bribes the emitter with between $0 and $1 to cut emission by one unit benefits both parties. Extension of this reasoning leads to the idea of an efficient level of emission, which negotiation should achieve, where the benefit to the emitter and the cost to the victim, for small changes in emission, are equal and the scope for further benefits to both parties from emissions reduction has been exhausted (efficient emission is illustrated in Fig 5 below).[iii]
The real situation with respect to GHG emissions represents the opposite extreme to this idealized example: rather than there being a single victim there are billions, and rather than the damage being contemporaneous, it will fall mostly on people who are yet to be born. Therefore public interventions such as taxes or quotas, rather than private negotiation, are required to reduce emissions to an efficient level. Since the externality spills across national boundaries, however, such coordinated action only at national level will not lead to sufficient emissions reduction: binding international agreement is essential for this. Moreover, in contrast to the idealized case of the preceding paragraph, it is impossible for the current losers from mitigation effort to be compensated by the future gainers, so social value judgements regarding the relative worth of gains and losses to different individuals and at different times are inescapable.[iv] The global and intertemporal nature of the GHG externality, together with its magnitude, underlie the comment in the Stern Review that this constitutes ‘the greatest market failure that the world has ever seen’. [Stern Review, p. xviii]
The economic theory of climate policy
This must start with the objectives of policy. The approach that is usual in Economics is both utilitarian and consequentialist. The utilitarian approach means that, ultimately, the costs and benefits of action are assessed by adding up estimated gains and losses to individual persons, now and in the future. These gains and losses can be expressed in monetary terms (how many dollars would compensate you for some climate change related damage?) or in ‘real’ terms, that is, what the required compensation would be in terms of some commodity or ‘basket’ of commodities. Consequentialism means that policy actions are judged solely in terms of their consequences, no account being taken, for instance, of the political process that leads to them.
The utilitarian-consequentialist framework dodges some issues and raises others. Among issues dodged, the most important is any notion that the environment should be valued as such, without reference to the sum of human utilities. According to the IPCC, 30-40% of all known species could become extinct by 2100 as a result of climate change: utilitarianism implies that this is of concern only to the extent that it is possible to trace consequences of this for individual well-being (Heal 2009). Among issues raised, two are paramount: income distribution and uncertainty.
As regards income distribution, both inter- and intra-generational: how should gains and losses to different individuals at the same or different times be valued relative to one another? The inter-generational question has two aspects to it. First, is there any justification for discriminating against persons simply because they belong to future generations? Such pure time preference[v] is hard to justify ethically, other than as a response to the possibility that some event will wipe out humanity before the future arrives. This consideration leads to positive, but very low, discounting[vi] of future gains, as discussed below (Stern 2007, p.31). Second, if we assume that the general pattern of rising prosperity during the past two hundred years will be replicated in the next two hundred, how should this be taken into account? In that case, mitigation efforts today will, on average, benefit individuals in the future who are better-off than those who bear the mitigation costs today. Therefore optimal mitigation depends inescapably on how society values extra units of consumption to richer and poorer people respectively, in other words to society’s degree of inequality aversion. If higher future prosperity is assumed and inequality aversion is positive, then gains in future consumption should be valued less than gains today, providing another reason for discounting, on top of any pure time preference. The higher is inequality aversion, the more future gains should be discounted and the lower mitigation should be in the present.
This line of reasoning, however, must be modified to take account of intra-generational inequality, for it is likely for several reasons that mitigation costs will fall on the relatively rich today, while the future benefits will accrue to the relatively poor. It is inevitable that the bulk of any mitigation now will occur in the richer countries of the world, imposing costs that will fall mainly on individuals in these countries. The two most important categories of future beneficiaries will be: (a) those whose livelihoods depend directly on agriculture in areas where agricultural production is particularly vulnerable to climate change, notably because of increased water stresses associated either with changes in rainfall patterns or glacier-fed river flow; (b) those living in areas vulnerable to sea-level rise and for whom relocation would be costly. Clearly these categories are largely made up of relatively poor people in developing countries. In sum, it is far from certain that future gainers from present mitigation will be any richer than current losers, in which case the argument via inequality aversion for discounting would evaporate.
Turning to uncertainty, analysis of its consequences for climate policy within the utilitarian framework begins from a presumption that individuals are risk-averse (and that policymakers’ decisions should respect this). At the individual level, risk-aversion is indicated by (indeed can be defined by) an unwillingness to take fair gambles. It seems a fairly secure generalization that, with respect to gambles that are large enough that the pleasure of gambling per se can be neglected, individuals do display considerable risk aversion. Few of us would stake the value of our house on the toss of a coin.
We have seen that forecasts of climate change itself are subject to a high degree of uncertainty. This uncertainty is greatly compounded by additional uncertainty over the human consequences of any given temperature rise. In general, and assuming risk-aversion, taking account of these uncertainties tends to shift optimal climate policy in the direction of higher mitigation now. Why?
Most analysts have answered this question by dividing it into two parts, making a distinction between what, for want of a better word, I will call ‘ordinary’ (Bell curve) uncertainty and catastrophe (see for example Stern 2007). For the first, suppose we are considering a choice between business-as-usual and a given programme of mitigation, which would stabilize atmospheric GHGs at some given level. To keep matters simple assume, as in Figure 1, that under mitigation future consumption (per head, say) is certain at CMIT. Under business-as-usual, suppose there is uncertainty regarding CBAU, indicated by the Bell curve probability distribution in the figure. Expected consumption is CEBAU, but risk aversion implies that the Bell curve outcome is worse than if BAU had entailed CEBAU for certain. Taking account of uncertainty therefore tilts the argument in favour of any given mitigation programme, so will clearly raise the optimal amount of mitigation.