Recently ‘insurance against risk’ has been proposed as a major addition to the list of economic supports for a high perceived value of children in poor South Asian countries. That is, that children provide a vital insurance policy for poor couples against the many hazards and risks to which they are subject. This argument is plausible and has proved appealing. If correct, it implies that demand for children will remain high (and contraceptive practice low) until the climate of risk has been substantially altered through public policy. This article examines this proposition critically with respect to the underlying logic and also subjects the proposition to historical and cross-sectional empirical test. We conclude that there are serious conceptual and theoretical weaknesses in the basic argument which cast serious doubt on its validity. Also the empirical tests undertaken lead us to reject the hypothesis. We conclude that the ‘risk’ notion adds nothing to fertility theory and that it remains to be established that comprehensive social security schemes are a necessary condition for fertility decline.