Demographic consequences of unpredictability in fertility outcomes.

Child survival is probabilistic, but the unpredictability in family formation and completed family size has been neglected in the fertility literature. In many societies, ending the family cycle with too few or too many surviving offspring entails serious social, economic, or fitness consequences. A model of risk- (or variance-) sensitive adaptive behavior that addresses long-term fertility outcomes is presented. The model shows that under conditions likely to be common, optimal, risk-sensitive reproductive strategies deviate systematically from the completed family size that would be expected if reproductive outcome is were predictable. This is termed the "variance compensation hypothesis." Variance compensation may be either positive or negative, resulting in augmented or diminished fertility. Which outcome obtained is a function of identifiable social, economic, and environmental factors. Through its effect on fertility behavior, variance compensation has a direct bearing on birth spacing and completed fertility, and thereby on problems in demography and human population biology ranging from demographic transitions to maternal depletion and child health. Risk-sensitive models will be a necessary component of a general theory of fertility.