Behavioural economics has shown that people are generally loss averse, and that reference points matter to how people perceive the impacts of policies[i]. Loss aversion posits that the negative feeling an individual experiences from a loss is larger than the positive feeling he/she experiences from a gain of the same size. In other words, the joy you feel from receiving a R500 lottery win is always going to be exceeded by the pain you feel from receiving a R500 speeding fine.

This phenomenon has been seen so often in experiments that it has led to one of the most well-known taglines of behavioural economics, namely “Losses loom larger than gains”[ii]. For the more technically minded, this leads a typical individual to have a kinked subjective value function that is steeper in the loss domain than in the gains domain as shown in Figure 1[iii].


Figure 1 Prospect Theory’s Subjective Value Function

Interestingly, the intensity of the feeling that someone experiences as a result of a gain or loss is not based on the size of the impact alone, but is influenced by the individual’s point of reference (which is often, but not always, a person’s net value at the point in time when the gain or loss occurs)[iv]. To illustrate this point, the utility that a poverty-stricken person experiences from receiving R10 is greater than the utility a wealthy person would experience from the same R10. This makes intuitive sense and could be explained by the difference in reference points these two individuals hold, i.e. their respective initial stocks of wealth.

However, the reference point an individual holds is not necessarily their initial endowment, it could be set in a number of ways, for example: what they think their net wealth is; what they expect their total future endowment to be; or even some bare-minimum level of wealth. The reference point could even be behaviour or endowment of others[v], so if you are given a bicycle and everyone else in society is also given a bicycle, you will be happier than if you are given a bicycle and everyone else is given a Porsche.

These two elements of loss aversion and reference dependence show that people do not always behave in the calculated way economists expect them to behave. Instead, psychological factors like loss aversion introduce systematic biases into our choices which can explain a host of seemingly irrational behaviour.

Companies’ strategic decisions are made by a collection of human beings who are influenced by systematic cognitive biases[vi]. It is thus possible that the behaviour of entire companies might also deviate from the predictions of rational choice theory, and the design of South African Carbon Tax seems to have taken this insight into consideration.

At its core, the carbon tax seeks to impose a cost on companies’ greenhouse gas emissions of R120/tonne. This tax rate is around the US $9/tonne mark using the average exchange rate for 2015 (R/$ 12.78), which places it in the same carbon price range as the European Union Emissions Trading System (EU ETS) and only slightly below the effective carbon price resulting from the Special Gas Emitters Regulation (SGER) in Alberta, Canada[vii]. This represents an impressive long-term commitment to tackling climate change and sharing the global responsibility of climate change mitigation. However, such a high rate may not be suitable to South Africa’s current economic conditions. Highlighting this is the fact that the effective carbon price in other middle-income nations, such as Mexico; Kazakhstan; and some Chinese Provinces, hovered below US $3/tonne in 2015[viii].

Therefore, in order to reduce the burden on South African companies, the carbon tax design incorporates a number of discounts that significantly reduce the effective tax rate to R48/tonne CO_2e (just over US $3 in 2015 prices) or lower (with the minimum carbon tax rate in some sectors being R6/tonne of CO_2e). The largest discount, which is applicable to all emissions, is a 60% basic allowance that reduces the effective carbon tax rate to the abovementioned R48/tonne of CO_2e.

Herein lies the wisdom of behavioural economics. The draft explanatory memorandum to the Draft Carbon Tax bill states that the discounts will be reduced over time to strengthen the carbon price signal. Assuming that the allowances will reduce over time until there are zero discounts, and the effective carbon tax rate is R120/tonne CO_2e, this would result in a tax rate that would be similar to the rest of the developed world’s tax rates, contingent on future exchange rates and changes in global carbon prices.

Consider the two ways in which the effective carbon tax rate could be communicated, namely the above-mentioned mechanism of reducing the 60% free basic allowance over time (ignoring the other discounts for the sake of simplicity), compared to increasing the tax rate gradually from zero. In the first scenario, setting the initially high tax rate and incorporating a gradually reducing free basic allowance (High Tax Basic Allowance (HTBA)) might potentially change companies’ reference points compared to if the tax rate was set lower, with no free basic allowance, and then raised each year (Low Tax No Basic Allowance (LTNBA) – the second scenario). This would be an example of framing – whereby two logically equivalent options are presented in semantically different ways[ix]. Under the HTBA frame, the intention seems to be to make R120/tonne the natural reference point, whereas under the LTNBA frame, the natural reference point would typically be R0/tonne (since there was no tax previously in place). Figure 2 illustrates the possibility of a shifted reference point under the HTBA frame compared to the LTNBA frame.

The world is a turbulent place; ripples from a huge range of global ‘events’ could have destabilising effects. How can private or public sector planners spot the ‘probabilities’ amid the fog created by the vast number of ‘possibilities’ – and, among them, those that could have a significant impact?

Figure 2 shows the possible changes in utility, represented by the progression of emoticons, associated with each carbon tax frame over time. The green bar in the HTBA framing represents the initial 60% tax-free basic allowance resulting in a R48/tonne effective carbon price. It is green to illustrate that this might be felt as a gain relative to the reference point of R120/tonne. Whereas in the LTNBA framing, the initial R48/tonne is red to show that it could be felt as a loss relative to the R0/tonne reference point. As time progresses from left to right, the effective carbon tax increases under both frames, either by decreasing the free basic allowance (HTBA) or by simply increasing the tax rate (LTNBA).

The HTBA framing of the carbon tax thus potentially induces less pain to, and hence more cooperation from, companies throughout the lifetime of the tax even though the total costs incurred under both frames are equivalent. The key is the R120/tonne reference point made salient in the HTBA framing. The gradual increase in the effective tax rate over time under HTBA is felt by companies as a gain that is being diminished; whereas under LTNBA the increasing tax can only be viewed as an increasing loss. This is illustrated in Figure 2 by the emoticon associated with the HTBA tax frame being significantly less angry than the corresponding emoticon associated with the LTNBA frame when time reaches the final effective tax rate of R120/tonne.

There are many other design elements to consider when looking at the Draft Carbon Tax Bill. However, in this respect it appears as though South African policy-makers have taken heed of the insights available from the field of behavioural economics to try increase the support for, or rather reduce the resistance to, the tax.

[i] Kahneman, D. & Tversky, A., 1979. Prospect Theory: An Analysis of Decision under Risk. Econometrica, 47(2), pp. 263-292.

[ii] Kahneman, D., 2011. Thinking Fast and Slow. New York: Farar, Straus and Ciroux.

[iii] Kahneman, D., Knetsch, J. L. & Thaler, R. H., 1991. Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias. The Journal of Economic Perspectives, 5(1), pp. 193-206.

[iv] Kahneman, D., Knetsch, J. L. & Thaler, R. H., 1991. Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias. The Journal of Economic Perspectives, 5(1), pp. 193-206. & Ariely, D., Huber, J. & Wertenbroch, K., 2005. When do losses loom larger than gains? Journal of Marketing Research, Volume 42, p. 134–138.

[v] Bogliacinoy, F. & Ortoleva, P., 2014. The Behavior of Others as a Reference Point. Columbia University: Working Paper.

[vi] Lovallo, D. & Sibony, O., 2010. The case for behavioral strategy: McKinsey & Company Quarterly.

[vii] Kossoy, A. et al., 2015. State and Trends of Carbon Pricing 2015, Washington D.C.: World Bank.

[viii] Kossoy, A. et al., 2015. State and Trends of Carbon Pricing 2015, Washington D.C.: World Bank.

[ix] McClure, J. & Sibley, C., 2011. Framing effects on disaster preparation: is negative framing more effective? The Australasian Journal of Disaster & Trauma Studies, Volume 2011-1.