In this post I am going to give a summary of the second part of Climate Economics: The State of the Art by Frank Ackerman and Elizabeth Stanton of the Stockholm Environment Institute, which deals with recent advances in the economics of climate change. Part 1, with a discussion of newest results from the climate science, can be found here. A summary of the above report’s overview of research in the economics of mitigation and adaptation will follow.
As discussed in my last post, recent findings in climate science indicate that climate change consequences are likely to be more severe, that the uncertainty with regard to the exact magnitude of change may be irreducible, and that our climate may be changing at a faster pace than previously thought. All these findings are highly important for the analysis of economic impacts of climate change and for the policy recommendations drawn from this analysis.
The authors of Climate Economics: The State of the Art identified three key issues for a modern economics of climate change to encompass: uncertainty (both with regard to the climate system – discussed here – and the economy), time spans (reflected mostly in discounting approaches) and the global nature of the climate problem (the latter two can be summarized also as equity issues). Since they are interrelated, I am going to connect the discussions of these problems with each other.
The Stern Review, published 5 years ago, triggered a general debate about the “right” shape of the economic analysis of climate change. Stern especially called for, on the one hand, a more accurate representation of the inherent uncertainty in the climate and economic systems (although his own model, PAGE, didn’t make a significant departure from common practice in this area), and, on the other hand, he used a social rate of discount that was controversially low for some other leading economists in the field. Economists haven’t yet reached consensus on these issues (and it is thinkable that they won’t ever agree), but the debate resulted in plentiful interesting approaches. Ackerman and Stanton discussed many of them, which I am going to summarize now.
As was discussed in my previous post, it is well possible that some sources of uncertainty with regard to the climate system (especially the climate sensitivity) cannot be removed. We have to live with them. However, many economists who have analysed the climate change problem have interpreted mean values from climate science as certain in their models (e.g., setting the climate sensitivity parameter equal to the IPCC’s so-called “best estimate”). There were some advances in the Stern Review (namely, a more extensive application of Monte Carlo analysis), but they appeared to lag behind the needs suggested by climate science.
An interesting departure in the area of uncertainty treatment was proposed by Martin Weitzman. In his mathematically complex analysis of the Dismal Theorem (I shortly discussed it here), he showed that, given the inherent uncertainty about the climate system and the non-negligible probability of a catastrophe, the prospect of this probable catastrophe dominates every analysis of the costs and benefits of climate change mitigation. This would effectively lead to an abandonment of the cost-benefit analysis and a switch toward a standards-based cost-effectiveness analysis – i.e., setting a “physical” goal (e.g., a temperature or greenhouse gas concentration threshold), and economically analysing which is the best way to achieve that goal. The assumptions of Weitzman’s analysis has been subject to technical criticism. Their technical implications may be problematic (e.g., if we assume that there is a danger of a climate catastrophe destroying civilization, as Weitzman originally did, one would expect an infinite willingness to pay for it to be averted), nonetheless they appear realistic and reasonable – the problem is rather the rationality assumption of current economics, which leads to such problematic implications as the above cited willingness-to-pay paradox.
Along with climate system dynamics, another important source of uncertainty in climate economics is the so-called damage function. In most integrated assessment models a quadratic damage function of temperature is assumed (i.e., a 2°C temperature rise would result in 4 additional units of damage) – a highly arbitrary solution to the problem of the complexity of possible climate impacts. Yet this ad-hoc assumption was made by most climate economists to date, including in Nicholas Stern’s PAGE model. However, a more thorough analysis of possible climate damages indicates that while the relation of temperature increases to damages may be well approximated by a quadratic function for lower temperatures, at higher ones the damage function should become much steeper. In the end, however, the damage function remains arbitrary to some extent simply because a more realistic projection of damages due to climate change exceeds the capacity of economic modelling.
Discounting is the most controversial and perhaps the most decisive issue in climate economics. Since I have already once discussed the issue quite extensively, I shall point to a few additional aspects from Ackerman and Stanton’s analysis:
- while it is clear that discounting is an ethical issue (descriptive approaches are not suited for the analysis of climate change, as I explained here), there are still many critical issues with regard to the Ramsey-equation, especially the question of the meaning of the parameter η and the proper interpretation of empirical data on people’s risk aversion (in the financial literature the problem is known as the “Equity Premium Puzzle” – an empirical and theoretical discussion of this can be read here);
- another issue related to the Ramsey-equation is that its mostly used form is an approximation – however, if the future rate of consumption growth is uncertain, it should have the more general form: ρ = δ + ηg – 1/2*η^2*σ^2 (where ρ is the discount rate, δ the rate of pure time preference, η the equity weighting parameter, g the growth rate of consumption, and σ the standard deviation of g), leading to lower values of the social rate of discount;
- some economists argue that it is not a reasonable assumption that people have the same (implicit) discount rate both for marketed goods and for the natural environment, and that the latter should be significantly lower (among others due to the non-substitutability of many livelihood-supporting functions of Nature);
- with regard to the descriptive approach to discounting, some argue that mitigation, as a risk neutral or even risk reducing activity, should be discounted at the rate of return on risk-free assets (e.g., government bonds).
Since the problems of choosing the right discounting approach and interpreting the parameters of the Ramsey-equation are inherent and probably unsolvable, some economists proposed alternative approaches to the problem of intergenerational impacts. One example is the overlapping generations (OLG) model, where in every time period two generations are assumed to coexist – the old and the young. Applied to the climate problem it mostly indicates benefits from rapid and decisive action to tackle climate change.
Due to the problematic issues of uncertainty and the treatment of intergenerational impacts, which appear to be inherent parts of cost-benefit analysis, there have been calls for a standards-based approach to be applied instead. Under this approach, a certain temperature-increase or greenhouse gas concentration target would be set (e.g., 2°C or 450 ppm CO2), and a cost-effectiveness analysis would be conducted to identify least-cost strategies to achieve this preset goal. In the next part of this discussion I am going to give an overview of Ackerman and Stanton’s analysis of the economics of mitigation and adaptation that is central to the standards-based approach.
After these strictly economic problems, Ackerman and Stanton also discuss some rather political issues related to global climate negotiations. An important question in global climate negotiations is how the effort should be allocated between countries (“burden sharing”). Different approaches are possible – I discussed here the Greenhouse Development Rights framework, which is politically problematic, although its ethical appeal is remarkable. Most proposals are more modest and call for allocation of remaining emission rights according to population and/or income. Further, climate negotiations can be analysed according to game theory. An interesting analysis is given in this working paper.
In their conclusion, Ackerman and Stanton give the following suggestions with regard to climate economics:
- Climate-economics models should use an up-tp-date representation of the climate system, including non-declining temperatures on a timescale of several centuries.
- Outcomes from climate change are uncertain, and climate-economics modeling results should reflect this uncertainty.
- Climate-economics modes should incorporate up-to-date scientific findings on the expected physical and ecological impacts of climate change.
- If damages cannot be accurately represented in welfare-optimization models, economics should instead use a standards-based approach.
- All climate-economics analyses should be accompanied by an explanation of what discount rate was chosen and why.
- Policy relevance in climate economics depends on the ability to present impacts not just for the world as a whole but also by region or income group.