Abstract The more health care is socialized, the more cost-effectiveness is an appropriate criterion for expenditure. Utility-maximizing individuals, facing divisibility of health care purchases and declining marginal health gains, and complete information about probable health improvements, should buy health care according to its cost-effectiveness. Absent these features, individual health spending will not be cost-effective; and in any case, differences in personal utilities and risk aversion will not lead to the same ranking of health care interventions for everyone. Private insurance frees consumers from concern for cost, which undermines cost-effectiveness, but lets them emphasize effectiveness, which favors value for money. This is most important for costly and cost-effective interventions, especially for poor people. Cost-effectiveness is more appropriate and easier to achieve under second-party insurance. More complete socialization of health care, via public finance, can yield greater efficiency by making insurance compulsory. Cost-effectiveness is also more attractive when taxpayers subsidize other's care: needs (effectiveness) take precedence over wants (utility). The gain in effectiveness may be greater, and the welfare loss from Pareto non-optimality smaller, in poor countries than in rich ones.