Stern's conclusions have been challenged strenuously (see eg Nordhaus, 2006 and Dasgupta, 2006). These challenges have largely been based on assumptions made about the present value of future costs made by the Stern Review. Assuming the expected utility model of prudential reasoning, Stern employs a discount rate for estimating the present value of the future costs of choices derived from Frank Ramsey, simplified by the assumption of a tractable growth of utility function. On this model of future costs and benefits, the discount rate r = d + h dc/dt,  where d is the measure of ‘pure time preference' (the discount from the worth of welfare in future times compared to find its present worth), h is the elasticity of consumption, and dc/dt is the per capita rate of growth of consumption. Some critics complain that Stern has assumed that d should be near zero, where only a tiny discount is made form the value of future welfare on the ground of the (remote) possibility that people become extinct. Stern also assumes that h = 1. Others complain that Stern thereby assumes that a 1% increase in future consumption represents (almost) the same increase in welfare as a 1% increase in present consumption. In part, these criticisms reflect a dispute within welfare theory, where confusion reigns about the properties and realism of models. But the dispute also reflects differences over fundamental ethical assumptions.

In setting d near zero, Stern is saying that we should value the welfare of someone in the future (nearly) as much as we value our own welfare. This is in accord with Rawls' (2001: ) principle of justice between generations. Claims that this is empirically not well founded since people discount their own future welfare, forget that there is a difference between the care one might take for one's future welfare, given mortality, and the care one should take for the welfare of future people, given only the possibility of extinction. Thus people in old age would be silly to take steps to provide for their own welfare 50 years hence, whereas the young quite sensibly make substantial sacrifices for their welfare 50 years hence.

Dispute over the appropriate value of h is much more arcane. Stern claims that setting it at unity is reasonable: it means that the welfare of an extra 1% for a rich person is equal to the welfare of an extra 1% for a poor person. This would justify transferring income from the rich to the poor until the cost of transfers brought down the gains in welfare for the poor to the point that there was a net loss of welfare from the transfer. Thus a transfer from a person earning $100,000 to a person earning $10,000 would be justified by a net gain in welfare if the cost of the transfer is less than $9000. This does not seem entirely unreasonable, given that justifications on utilitarian grounds of enhancement of net welfare may not give all the reasons we have for making income distribution more equitable.[9]

In any event, Stern estimate for the discount rate r is just above the growth rate of consumption. This brings it close to the real risk free bond rate, which seems a more appropriate reference point for the discounting long term future prospects than the real rate of return on investment projects of 5-6%. We conclude that Stern's estimates of the costs of benefits of global warming, while contentious, are sound enough for precautionary reasoning[10].