Using a simple societal utility function--giving equal weight to current and future generations-it is concluded that costs need to be discounted on the basis of the expected increase in income and the marginal utility of consumption, and that effects need to be discounted on the basis of the expected increase in health and the marginal utility of health. It is derived that both rates need to be equal when assuming a kind of perfect market, where growth rates are determined by the societal utility function. It is argued that this is an extremely heroic assumption and that different discount rates may be needed. Additionally, the traditional 'inconsistency arguments' of Weinstein and Stason and of Keeler and Cretin are reconsidered. Within the context presented earlier, the first inconsistency only emerges when a growth equilibrium is assumed, reinforcing the arguments put forward before. The Keeler and Cretin paradox is reconsidered by showing that absolutely no paradox emerges when programs are not supposed to stop after a year but are supposed to continue indefinitely. The conclusion is drawn that non-believers in market mechanisms assuring an optimal social policy, need to reconsider the use of their discount rates.