Declining Discount Rates: The Long and theShort of itBEN GROOM1;, CAMERON HEPBURN2, PHOEBE KOUNDOURI3,and DAVID PEARCE41Department of Economics, School of Oriental and African Studies, London, UK;2St Hugh’sCollege, Environmental Change Institute and Department of Economics, Oxford University, UK;3Department of Business, Department of Economics, Reading University, UK;4Department ofEconomics, University College London, UK;Author for correspondence (e-mail: [email protected])Accepted 14 March 2005Abstract. The last few years have witnessed important advances in our understanding of timepreference and social discounting. In particular, several rationales for the use of time-varyingsocial discount rates have emerged. These rationales range from the ad hoc to the formal, withsome founded solely in economic theory while others reflect principles of intergenerationalequity. While these advances are to be applauded, the practitioner is left with a confusingarray of rationales and the sense that almost any discount rate can be justified. This paperdraws together these different strands and provides a critical review of past and presentcontributions to this literature. In addition to this we highlight some of the problems withemploying DDRs in the decision-making process, the most pressing of which may be timeinconsistency. We clarify their practical implications, and potential pitfalls, of the morecredible rationales and argue that some approaches popular in environmental economicsliterature are ill-conceived. Finally, we illustrate the impact of different approaches byexamining global warming and nuclear power investment. This includes an application andextension of Newell and Pizer [‘Discounting the benefits of climate change mitigation: howmuch do uncertain rates increase valuations?’ Journal of Environmental Economics andManagement 46 (2003) 52] to UK interest rate data.Key words: global warming, intergenerational equity, social cost benefit analysis, timeinconsistency, uncertainty, time varying discount rates1. IntroductionDebates about discounting have always occupied an important place inenvironmental policy and economics. Like all other investments, investmentin the environment involves incurring costs today for benefits in the future.Whether a public investment is efficient or not is determined by social costbenefit analysis (CBA). Where welfare is Utilitarian, the socially efficient levelof investment is attained by investing in projects where the net present valueEnvironmental & Resource Economics (2005) 32: 445–493 Ó Springer 2005DOI 10.1007/s10640-005-4681-y(NPV), determined by discounting costs and benefits at the social discountrate (SDR) over the time horizon, is greater than zero.1It follows that thelevel of the SDR is critical in determining whether an individual publicinvestment or policy will pass a CBA test.Quite separately from arguments over whether the discount rate shouldbe positive or not (e.g., Olson and Bailey 1981; Broome 1992), economistsand others have argued at length over which of several potential discountrates should be used as the SDR (e.g., Marglin 1963; Baumol 1968; Lind1982). Several candidates exist, the most widely recognised of which are thesocial rate of return on investment (r) and the rate at which society valuesconsumption at different points of time, the Social Rate of Time Preference(d). The distinction between these discount rates is most important in thesecond best world in which distortions to the economy, such as corporateand personal taxes or environmental externalities, prevent these rates frombeing equalised. The choice of SDR is inherently complicated in such sit-uations and is dependent upon a wide variety of factors. These factorsinclude: the extent to which public investment displaces or generatesconsumption or private investment throughout the lifetime of the project,the extent to which project risk is captured by the discount rate, andassumptions concerning reinvestment (Lind 1982; Portney and Weyant1999). However, one thing common to much of the past literature is that,whatever the rate chosen, the relative weights applied to all adjacent timeperiods would be invariant across the time horizon considered. That is,discounting would be exponential.A common criticism of discounting is that it militates against solutionsto long-run environmental problems: for example, climate change, biodi-versity loss and nuclear waste, which need to be evaluated over a timehorizon of several hundred years. The question arises: What is the appro-priate procedure for such long time horizons? There is wide agreement thatdiscounting at a constant positive rate in these circumstances is problem-atic, irrespective of the particular discount rate employed. With a constantrate, the costs and benefits accruing to generations in the distant futureappear relatively unimportant in present value terms. Hence, decisionsmade today on the basis of CBA appear to tyrannise future generations andin extreme cases leave them exposed to potentially catastrophic conse-quences. Such risks can either result from current actions, where futurecosts carry no weight, e.g. nuclear decommission, or from current inaction,where the future benefits carry no weight, e.g. climate change. The inter-generational issues associated with discounting have puzzled generations ofeconomists. Pigou (1932) referred to the deleterious effects of exponentialdiscounting on future welfare as a ‘defective telescopic faculty’. More re-cently Weitzman (1998) summarises this puzzle succinctly when he states:BEN GROOM ET AL.446‘to think about the distant future in terms of standard discounting is tohave an uneasy intuitive feeling that something is wrong, somewhere’.Discounting also appears to be contrary to the widely supported goal of‘sustainability’, which by most definitions implies that policies and invest-ments should contribute to securing sustained increases in per capita welfarefor future generations (Atkinson et al. 1997). Also, by attaching little weightto future welfare conventional discounting appears to ignore any notion ofintergenerational equity.A recently proposed solution to this problem is to use a discount rate,which declines with time, according to some predetermined trajectory, thusraising the weight attached to the welfare of future generations. It is possiblethat using a declining discount rate (DDR) could make an important con-tribution