Numerous authors have proposed schemes whereby delayed effects may be discounted in relation to early effects for both societal and individual risks of fatality. The principal application of such proposals is in the field of major hazards risk assessment. This paper reviews the literature on this topic and identifies numerous such discount factors. The basis of each is discussed, and a scheme is proposed to enable the quantitative arguments to be explored. The principle employed is that the rate of loss of life expectancy should be the same for delayed effects as for early effects. This is applied to representative populations in the context of societal risk limit lines. The scheme is generalized to encompass various forms of such criteria. Some clarification is offered concerning identified problems of interpretation arising from the literature on societal risk criteria. It is concluded that, except for very long induction periods between exposure and manifestation of effect, the application of such discount factors can only be justified where exceptionally high accuracy is demonstrated in the risk analysis, and that usually the analysis uncertainties are such as to preclude the use of such discounting. Additionally, in consequence of these considerations, it is recommended that delayed and early deaths should be treated identically for the purposes of risk assessment.