Discounting the distant future has periodically been a controversial topic in welfare economics but the evaluation of climate change policy and particularly the Stern Review have given the debate a new relevance. The parameters in a standard social welfare function that determine the path for the discount rate are also important in determining the time path of saving, and several prominent economists have criticised the values used in the Review specifically because they imply excessively high optimal saving rates, from either a positive or normative perspective. The fact that near-zero rates of pure time preference do not necessarily lead to absurdly high saving rates has been known for some time. However, in the context of climate change policy, this point has been made using inappropriate models or specific numerical examples with a rather arbitrary value for the rate of growth of total factor productivity (TFP). Given the attention that the ‘unreasonable saving rates’ debate has received in the climate change literature, there is a role for a rigorous presentation of the determinants of saving rates in models used to evaluate climate change policy, using values for TFP growth informed by recent historical experience. I show that both in theory and practice, optimal saving rates in the presence of near-zero pure time preference are far from the near-100 per cent ones obtained from simpler models. In the widely used Dynamic Integrated model of Climate and the Economy (DICE) model, optimal rates are close to 30 per cent for a range of values of the elasticity of marginal utility of consumption, and for Stern’s revised central value for that parameter they do not exceed 31 per cent. While the role of TFP growth in lowering optimal saving rates in the presence of near-zero rates of pure time preference may have been overplayed in some previous work, TFP growth is a key determinant of output and hence emissions and climate damage, so working with realistic values of TFP growth remains crucial.