66 Cost Benefit Analysis

Spash on Dasgupta Review

Decades ago environmental CBA developed a range of methods for imputing monetary values, but with limited applicability under specific conditions (Hanley & Spash, 1993; Spash, 2005). For a start, these methods only apply to marginal changes in environmental goods or services, not least because the value of money itself (its marginal utility) alters when there are large changes affecting income; also, economic welfare measures assume other things (e.g. all other prices) remain the same which is violated by large changes. Clearly things like mass extinction of species and human induced climate change are not small, marginal, changes. Two approaches are employed by social CBA: revealed and stated preference methods. The first relies upon existing markets that can be associated with environmental attributes (e.g. air pollution affecting house prices) and so is severely restricted. The second uses surveys designed to illicit, primarily, willingness-to-pay for environmental changes. Dasgupta attends to one stated preference approach: the contingent valuation method (CVM). He claims that: ‘CVM is attractive because it appeals to our democratic instinct, that people should be asked for their opinion on matters that may be of concern to them’ (Dasgupta, 2021, p. 304). In addition, the CVM is promoted as widely applicable (unbounded by existing markets) to revealing values for everything from aesthetics to biodiversity loss. The ability to include a range of value categories contributing to an individual’s utility extends to including: ‘respondents’ sense of a species’ existence value—perhaps even its intrinsic value’ (Dasgupta, 2021, p. 304). Normally environmental economists define a set of four values: direct use, option, existence and bequest value. However, Dasgupta (Dasgupta, 2021, p. 301), for no apparent reason, claims a different set of six sources of value for biodiversity that mix-up objects of value with types of values. While all his examples are consequentialist and based on creating utility for humans, he confuses concepts of existence value with sacred values, moral worth and intrinsic value. So let us turn to the value problems.

Economic welfare theory requires that people should be compensated, not pay, for an environmentally degrading imposition on them (e.g. pollution, biodiversity loss). However, this is generally not undertaken because people could ask large sums, destroying the economic calculus, so economists prefer to restrict respondents replies to their income (i.e. ability to pay) regardless of their own theoretical requirements for validity. Besides being the incorrect measure, willingness-to-pay is not a democratic approach seeking an opinion, as claimed by Dasgupta. Despite their attempts to control respondents, CVM produces results deemed unacceptable because people appear willing-to-pay too much or refuse to bid (i.e. protest).

Attempts have then been made to redesign the surveys to get the responses economists want and hence they developed choice experiments, cited as a problem solving advance by Dasgupta (2021, p. 304). Here respondents have restricted ability to protest, or violate the economists model of how they should behave, and can only refuse to answer completely and fall into the ignored nonrespondent category. Failing ‘success’ with survey design, collected data may be subject to manipulation to get the desired values.

The welfare economics underlying social CBA assumes that all values can be reduced to individual preferences as expressions of utility.

Refusing to make trade-offs is also a principled positon disallowed by mainstream economics or treated as an anomaly.

The valuation question of social CBA, ‘what is your maximum willingness-to-pay for more/less X?’, implicitly assumes there is no moral objection to the question itself.

More frogs do not equate to fewer tigers. What then is ‘natural capital’?

Health (mortality/morbidity) as a capital investment is even worse. Producing money numbers here requires the conjuring trick of talking about abstracted non-real people who are represented as ‘statistical lives’, under the VSL. For example, the results are used in transportation assessment to decide upon road building programmes and the installation of safety equipment. However, the public rejection of this approach is exposed when there is a train crash, people are killed and the public discover the lack of safety equipment is due to the calculation that it cost more than the expected fatalities times the VSL. Politicians rarely defend the numbers in such circumstances, although their transport departments may continue to use them on a daily basis.

IPCC

A major example of the failings of VSL arose during the third assessment report of the Intergovernmental Panel on Climate Change (IPCC). Willingness-to-pay informed VSL, based on Fankhauser (1995, p. 47), gave a range from $0.2–$16.0 million with an average of $3 million, and $1.5 million adopted as the VSL for developed countries. Adjustment was made for income to give ‘an arbitrary value of $300,000 for middle income and $100,000 for low income countries’. The result was a factor of fifteen difference between VSL in high ($1.5 million) and low ($0.1 million) income countries. A storm raged when the IPCC chapter employing this approach appeared (see Spash, 2002a, Chapter 7). Representatives from industrially developing nations, led by India and China, refused to accept the report citing it as absurd, discriminatory, unethical, technically inaccurate and anti the poor.

Shortly after the IPCC VSL controversy a prime example of commensurability problems arose when CBA was applied to climate change by Nordhaus (1998a, 1998b). He claimed increased moridity/mortality would be outweighed due to increased leisure opportunities by a factor of 30 to 10 in China and by 38 to 3 in the USA. An example Nordhaus was using at the time concerned claiming that golfers may view global warming as a boon to year-round recreation. So, if we extend this logic to global studies, more golfing days in Florida could compensate for dead people in China. The commensuration of values in The Review is no different. Classes of capital are values, equated and summed. Human capital is an aggregation of values so that, for example, more ‘education’ can compensate for increased risk of death. More than this, if education pays better financial dividends than avoiding loss of life then, according to the economic accountants, the optimal world should have more education and more death.

66.1 Discounting

The mechanism chosen by Dasgupta to allocate natural resources between current and future generations is a social discount rate (SDR). Alternatives, such as allocation on grounds of justice, rights or needs, are therefore excluded,

While acknowledging the major ethical objections to discriminating against future generations via discounting, Dasgupta (2021, Chapter 10) nevertheless chooses a positive discount rate. He justifies this using the dubious argument that since returns on judiciously chosen investments are positive (by assumption), fairness requires discriminating against future generations, because otherwise the current generation would be unduly limited in consumption and condemned to excessive poverty. This is a productivist logic based on assuming the future is always better-off.

Dasgupta follows standard neoclassical theory in treating future outcomes (flows of costs and benefits) for public policy projects as subject to a social time preference (STP). The basic positon here is to follow a formulae, called the Ramsey rule, that determines the SDR and STP as follows:

\[SDR = r + h g = STP\]

where g is annual per capita growth of consumption, η is the elasticity of marginal utility of consumption and ρ is the utility discount rate, consisting of a component for pure time preference, δ, and, in HM Treasury practice, a component for certain types of risk, L. Components of the formulae are so uncertain that economists appeal to surveying themselves to get estimates, as if this provided objective data.

In The Review commitment to an actual number is vague, and subject to speculation as to economic growth and uncertainty. Elsewhere Dasgupta (2008) has argued that δ could be zero, while, contrary to others, he argues for a much higher η in the range 2–3 or more. The basic rate used by HM Treasury, in its Green Book, is 3.5%, 6 where δ = 0.5, L = 1 so that ρ = 1.5 with the remainder consisting of consumption growth g = 2 and η = 1. If Dasgupta’s argument for η is adopted then the STP would be between 5.5% and 7.5%, but he has argued favouring δ = 0 which would give 5.0–7.0%. This is extremely high.

For comparison consider how Nordhaus uses such rates to recommend catastrophic global warming as economically rational. He states that ‘the cost-benefit optimum rises to over 3°C in 2100’ (Nordhaus, 2018, p. 452), and his Figure 5 shows a 2100 optimum around 3.6°C and rising, because he recommends discounting the future at around 5% (Nordhaus, 2018, p. 455), writing off any importance of future damages by 2100 (damages would weigh around 2% of their value today meaning, for example, under VSL an action saving 2 people today at the cost of killing 97 people in 2100 would be a net gain, an optimal choice). This is neither unique to Nordhaus nor new, for example, the economic working group of the IPCC third assessment used discount rates between 5% and 12%. A common claim, also made by Nordhaus, is that empirically observable rates of return should be used. However, in actual economies the rate of return on risk free investment has been zero or negative in real terms for years (Freeman et al., 2018, p. 16), but discounting has persisted regardless of the theoretical justifications.

In fact there is no such thing as a singular rate in actual economies. The recognition that differential rates for different projects is theoretically justified, and specifically for projects with environmental impact, has led HM Treasury to discount at a lower rate (1.5%) for project impacts on health and life. However, while recognized as formally correct Dasgupta rejects this on the grounds, not of theory but, that it will be ‘cumbersome’ in practice and ‘lead inevitably to errors’. Indeed he states it would be ‘unsafe’ because (now) the ‘social evaluator’ (or ‘citizen investor’) cannot be trusted to get things right. Instead he recommends a single rate applied to all projects. The reader might wonder at such pragmatism, for if all the problems of neoclassical economics can be so easily dismissed on a whim as impractical and too cumbersome for his social/citizen evaluator/investor this rather begs the question why he bothers us with his models, theories and extensive mathematical detours and what other fallibilities his central decisionmaker might suffer from. That markets are not guides to intergenerational fairness, ethics or equity, would seem to bring the whole approach into question. In his paper on discounting for climate change Dasgupta concludes: ‘Intergenerational welfare economics raises more questions than it is able to answer satisfactorily’. However, as usual, Dasgupta’s recognition of the problems has no impact on his esteem for and continued use of neoclassical economics and, unsurprisingly, he recommends discounting on this basis.

Spash (2021) The Dasgupta Review deconstructed: an exposé of biodiversity economics (pdf)