Prospect theory, mitigation and adaptation to climate change

Abstract

Prospect theory (PT) is a widely accepted theory for decisions under uncertainty. However, so far a systematic application to climate policy (CP) does not exist. One important postulation of PT is that outcomes are perceived as gains or losses, relative to the reference point. When it comes to CP, different decision-makers may have different reference points. For example, one decision-maker perceives the current climate as the reference point whereas another decision-maker may have another one, say climate in 100 years. For the former, climate damages are losses and the benefits of CP are reductions of losses. For the latter, benefits of CP are gains. PT suggests that the former places a higher value in CP than the latter. After a critical review whether and how PT may be applied to CP, the paper systematically presents this and other cases where PT offers new insights into climate-related analyses, notwithstanding the importance of well-known aspects such as discounting, altruism, political and economic costs. It is shown that accounting for PT may contribute to a better understanding of some well-known puzzles in the climate debate, including different preferences for CP amongst individuals and nations, the role of technical vs. financial adaptation, and the apparent preference for hard protection measures in coastal adaptation. Finally, concrete possibilities for empirical research on these effects are proposed.

Keywords:

• climate change
• prospect theory
• reference dependence
• adaptation
• mitigation

Acknowledgements

I would like to thank Klaus Eisenack, Leena Karrasch, Thomas Klenke, Anna Pechan, and Heinz Welsch from the University Oldenburg as well as Claudio Baccianti, Peter Heindl and Michael Schymura from the Centre for European Economic Research in Mannheim for useful comments and fruitful discussions.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. In this paper, the term ‘response measure’ refers to any private or public activity which is designed to either reduce greenhouse gas emissions (mitigation) or reduce the damage, given that climate change is occurring (adaptation to damages), or yielding the potential benefits of climate change (adaptation to benefits).

2. Contrary to the conclusions of the mentioned studies, Glenk and Colombo (2013) ‘cautiously advocate the use of a non-linear EUT model’ for modelling outcome-related risk evaluations in a choice experiment.

3. It should be noted that various generalizations of the conventional EUT as introduced by Von Neumann and Morgenstern (1947) exist. E.g. Quiggin (1982, 1993) proposed non-linear probability weighting. Numerous empirical works in the field of environmental economics base on the conventional or derived forms of EUT (e.g. Farsi 2010; Glenk and Colombo 2013 to name just a few). However, PT has been broadly accepted as the behavioural decision theory which is best in describing observed decision-making processes.

4. Note that probability weighting includes more than certainty effects. For example, the overweighting of small probabilities can be used for explaining the attractiveness of lotteries and insurance contracts (Tversky and Kahneman 1992, 316). However, this phenomenon is already well analysed (e.g. Botzen and van den Bergh 2009). In the following, we limit ourselves to the certainty effect implied by probability weighting because there we see applications to climate response measures which have not been discussed yet.

5. Though, unintended (‘incidental’) adaptation also exists, as discussed by Eisenack and Stecker (2012).

6. According to a review of empirical applications of PT, public finance and macroeconomics are topics where works based on PT are still relatively rare (Barberis 2013, 190). Arguably, one important reason is that decisions in these fields are often collective decisions.

7. Hallegatte (2013). follows this approach in his study on the optimal protection of economic capital in the light of natural disasters by referring to suboptimal decision-making and its impact on the optimal protection policy.

8. Similar to the concept of reference points in PT, the regulatory focus theory from psychological literature (e.g. Higgins 1998) distinguishes between people with a promotion focus who are predominantly engaged in approaching pleasure (gain domain) and people with a prevention focus who are rather focussed on avoiding pain (loss domain). By this differentiation, the theory provides one possible concept how different reference points may emerge on the individual level.

9. Note that also unexpected events may shift the reference point. Think of the situation when a tipping point in the earth’s climate system is reached and large-scale impacts become apparent. This may change the reference point of decision-makers.

10. Since decisions on the mix between adaptation and mitigation are presumably collective or institutional decisions, the application of PT is questionable (see Section 3).

11. This result depends on the concrete parameterisation of the value function. Here the widely used parameterisation of Tversky and Kahneman (1992) is assumed. Additionally, the adaptation costs have to be somewhat lower than the benefits.

12. Although insurance is broadly seen as an efficient adaptation to unpredictable damage events, the demand for insurance coverage is relatively low, especially for low probability/high impact events (Aakre et al. 2010; Schwarze et al. 2011). Even in the case of subsidised premiums, the demand remains remarkably low. This has been studied by numerous authors, and several explanations have been proposed. One main argument is that small probability events are overall ignored or potential damages are broadly underestimated by decision-makers (Kunreuther 1984; Kunreuther et al. 1978). Other explanations argue that the expectation of emergency relief is responsible for a low private demand (Raschky and Weck-Hannemann 2007).

13. Anecdotic evidence for this can be found in policy papers and reports (Doswald and Osti 2011, 23; Klimaschutz 2012, 51–55). Evidence that soft protection measures imply higher uncertainty than hard protection is apparent in cost-benefit analyses (Brouwer and van Ek 2004; Meyer, Priest, and Kuhlicke 2012). This uncertainty also hampers public acceptance (Weisner and Schernewski 2013).

14. Index-based insurances are frequently discussed as an appropriate strategy to adapt to increasing climatic variability (Herweijer, Ranger, and Ward 2009; Hochrainer, Mechler, and Pflug 2009; Linnerooth-Bayer and Mechler 2006).

15. Harrison and Rutström (2009) report a lower probability of PT-like behaviour for white men in their sample.

16. This is, however, subject to the open debate whether collective action is influenced by reference dependence (see Section 3).

17. Instead finding the predominant reference point within the population of a country, an obvious first step would be an analysis how policy-makers or climate negotiators of the given country define the climate reference. In autocratic regimes, this would be even more appropriate than the public perception.

18. It should be noted that the use of framing effects and nudging in general is sometimes debated as ethically problematic, because people’s choices are manipulated (Hausman and Welch 2010).