Ethics and Discounting Global Warming Damages
October 30, 2009
One of the most contentious issues in assessing what price to put on greenhouse gas emissions is the rate at which global warming damages to future generations should be discounted. How can we think about the discount rate and is there any possibility for reconciling different views?
Over the past few years, great debates have erupted over discounting, stimulated by the economic analysis of climate change. One of the more controversial, the 2006 review by Nicholas Stern, The Economics of Climate Change: The Stern Review, employed a much lower discount rate—around 1.4 percent—than had previously been used. Partly for this reason, the Stern Review recommended more rapid reductions in greenhouse gas emissions, with carbon dioxide (CO2) concentrations peaking at 450–550 ppmv (parts per million by volume). Stern’s implicit carbon price along a business-as-usual pathway was roughly U.S. $85 per ton of CO2, though the damages from today’s emissions (and hence the appropriate price on CO2) would be substantially lower if atmospheric stabilization targets were achieved.
In contrast, much (but not all) previous economic analysis had used market interest rates of well above 2 percent, with concentrations reaching around 700 ppmv and carbon prices of considerably less than U.S. $20 per tonne of CO2, and often single digits. By way of comparison, current carbon prices in the European greenhouse gas emissions trading scheme are around U.S. $20 per ton, and have been as high as around U.S. $40 per ton.
The discounting debate is certainly a critical issue for all of the economists now engaged in climate change policy discussions. But before engaging with the debate, it is helpful to clarify some key ideas. To start with, a discount rate is a rate of change of the price of one good relative to another. Under idealized circumstances, the discount rate for all goods could be identical. But the world is not ideal, so different goods have different appropriate discount rates. So when economists refer to “the” discount rate, we are referring to a general, economywide discount rate, which can roughly be applied to aggregate consumption in the economy. This general discount rate represents our collective willingness to trade off aggregate present for future consumption. The discount rate reflects changes in real (not just nominal) prices, and is not merely an adjustment for inflation.
Economists make an important distinction between the discount rate for consumption versus utility (or wellbeing). We might discount expected utility in the future because, for instance, there is a risk of dying beforehand. We might discount future generations because we care less about their welfare than we care about our own. In addition to discounting utility, a further discount component is applied to future consumption, if higher living standards are anticipated in future. This reflects the rate at which the value of additional consumption declines as consumption increases—represented by a parameter called the elasticity of marginal utility with respect to consumption. In total, the consumption discount rate comprises two parts: the utility discount rate and the elasticity of marginal utility multiplied by the consumption growth rate. Even with a zero utility discount rate, if aggregate consumption is expected to keep growing, then a positive consumption discount rate is appropriate.
The discount rate is a function of how we expect consumption to change in the future. Greater optimism (pessimism) about future consumption growth implies a higher (lower) consumption discount rate. In rare cases, where large-scale investment changes consumption growth, the investment will also change the appropriate consumption discount rate. Climate change, and/or our response to it, may be large enough to change the underlying growth rate, and hence also the discount rate.
So why is there such a difference between the Stern Review and most previous research? Stern adopts a “prescriptive” approach, explicitly considering the ethics of climate change, and his modelling treats the utility of everyone equally: individuals are not discounted just because they are born in the future. The average global per-capita consumption growth was set at around 1.3 percent (although consumption growth varies across regions and from each of the many thousand model runs to the next). With an assumed “elasticity of marginal utility” equal to one, this implied a global average consumption discount rate of 1.4 percent. As noted, Stern’s carbon prices were higher than in earlier work, and his recommendation is to reduce emissions rapidly.
In contrast to Stern’s prescriptive approach, previous research tended to be “descriptive” in assumptions about discounting, focusing on what we actually do, rather than what we ought to do from an ethical point of view. The focus was on market interest rates, which reflect the sum of many individual choices. Historic market interest rates (ignoring past and present financial crises) have averaged around 6 percent, so most previous research applied consumption discount rates at roughly this level. As such, utility discount rates were around 1–3 percent, the elasticity of marginal utility set at 1–2 percent and consumption growth rates of around 2 percent. With these higher discount rates, much more gradual emissions reductions are recommended, with atmospheric concentrations reaching or exceeding 700 ppmv.
Too stern about Stern?
Several arguments have been advanced against Stern’s approach. We consider two of the more powerful. Stern’s utilitarian ethics is not the only, or even the predominant, ethical outlook. For instance, “agent-relative” ethical ideas advanced by philosopher David Hume in the 18th century suggest that it is legitimate to care about those closer to us (by genetic proximity, or space or time) than those further away. Also, if our ancestors had adopted Stern’s perspective, they would have had to devote more resources (by way of savings and investment) for our benefit, reducing their own consumption, and hence also their welfare. This seems unfair given that our ancestors were significantly poorer than we are today.
There are two corresponding replies. First, it is true that utilitarianism is not the only viewpoint. But for a global issue like climate change, our analysis should be impartial, not favouring the Chinese over the Americans, say, or people alive today ahead of those alive in 2050. Many of the greatest economists and philosophers have specifically endorsed an impartial approach, and recommended a zero utility discount rate, which implies a low consumption discount rate, as in the Stern Review. Furthermore, the “prescriptive” school argued that there are at least three reasons for being cautious about using market prices to reveal ethical attitudes: markets do not always work properly, as we have seen in recent years; market interest rates can only aggregate the choices of those alive today, and do not represent the wishes of future generations; and interest rates reveal only one discount rate—yet, as discussed, in the real world different goods will have different discount rates.
Second, lower discount rates would indeed have compelled our ancestors to have saved more, but not ruinously so, once we account for the fact that we are so much better off than our ancestors were (as a result of impressive technological progress).
So, what to do?
There are several pragmatic routes to reconciling the approaches. One is to explicitly take into account uncertainty over the discount rate. For example, suppose we crudely assume that Stern is as likely to be correct as his critics, such that a 1.4 percent discount rate is as likely to be correct as a 6 percent discount rate. In this case, over a 100-year period, the discount rate that yields global warming damages equal to the average of damages under a 1.4 percent rate and under a 6 percent rate is 2 percent. In other words, the logic of uncertainty makes the arguments about ethics less significant.
Accounting for “unknown unknowns,” by assuming the probabilities are themselves uncertain, further bridges the divide between the discount rate of the Stern Review and higher market interest rates. Also, it may be consistent to apply high discount rates for aggregate consumption, and low discount rates for (increasingly scarce) natural capital. If climate change disproportionately damages natural capital, a lower discount rate may be justified.
Finally, social institutions, beyond the market for government bonds, might be investigated with a view to backing out implicit long-term ethical preferences. This would avoid the need for a priori ethical assumptions, such as those made in the Stern Review, and would also avoid the problems with relying on market prices to derive ethical positions. Such research may reveal that the apparent chasm between prescriptive and descriptive approaches is narrower than it seems.
Cameron Hepburn is a senior research fellow at the Smith School of Enterprise and the Environment at the University of Oxford. He specializes in environmental economics, climate policy, and long-term decisionmaking.