Abstract There has been considerable debate in the literature regarding the social security-fertility hypothesis. By analyzing 81 countries for which data was available, numerous econometric models were developed to test both the relative importance of social security as a causal factor affecting fertility and the subsequent effect of lower fertility levels on social security expenditures. By controlling for economic development and testing the model at different points in time, the social security variable was found to be causally linked to subsequent fertility levels as well as fertility levels being causally linked to subsequent social security expenditures. Furthermore, the appropriate response lag was identified and various nonlinear functional forms were utilized to control for economic development. Finally, a Chow test was employed to show the existence of structural change between the developed and less developed nations.