Barro's fertility equations: the robustness of the role of female education and income

Barro and Lee (1994) and Barro and Sala-i-Martin (1995) find that real per-capita GDP and both male and female education have important effects on fertility in their cross-country empirical studies. In order to assess the robustness of their results, their estimated models are subjected to specification and diagnostic testing, the effects on the model of using the improved Barro and Lee (1996) cross-country data on educational attainment of the population aged 15 and over are examined, and the different specifications used by Barro and Lee and by Barro and Sala-i-Martin compared. The results obtained suggest that their fertility equations do not perform well in terms of diagnostic testing, and are very sensitive to the use of different vintages of the educational attainment proxies and of the Summers-Heston cross-country income data. A robust explanation of fertility, to link with empirical growth equations, has, therefore, not yet been found; further work is required in this area.