Investigating the causality between female labor force participation and fertility, Turkey 1968-2006

Ayse Abbasoglu Ozgoren, Hacettepe University
Ahmet Sinan Turkyilmaz, Hacettepe University

The classical literature on female work force participation and fertility suggests an inverse relationship between the two, for developed countries. In contrast, evidence from the developing countries shows an ambiguous picture and, in general, fertility declines in these countries have been rather slower than those of developed countries, if not negligible. In Turkey, fertility rates have been declining accompanied with decreasing female labour force participation (LFP) rates. However it is not possible to determine the way of relationship or causality between fertility and working among women by just looking at the trends since there are more mechanisms affecting each variable in various ways. To the authors’ knowledge, no macro-level studies have yet been undertaken on Turkey to investigate the causal link between female LFP and fertility. This study applies the Johansen-Juselius Maximum Likelihood Error Correction Modelling, in conjunction with the ‘general to specific modelling’, for examining the existence of long-run relationship as well as the causal link between female LFP and fertility in Turkey using macro-level data over the period: 1968-2006. The specified multivariate model, consisting of total fertility rate (TFR), female LFP, infant mortality rate, and female illiteracy rate, is indicative of the existence of an inverse long-run relationship between fertility and female LFP. All the long-run estimates of other variables have the theoretically anticipated signs. LFP is negatively related to fertility. The elasticity of the fertility rate is around -0.5, suggesting that a 1% increase in female LFP rate reduced fertility by 0.5%. This finding supports the “role incompatibility theory”. The direction of causality is two-way for fertility and female labour force participation, i.e. there is a feedback relationship between TFR and female LFP rate. The causality exists both in the long-run and the short-run provided that infant mortality and female illiteracy are also included in the model.