Inverse or U-shaped educational gradient in fertility differentials? Evidence from census-linked data for Lithuania

Aiva Jasilioniene, Max-Planck Institute for Demographic Research
Vlada Stankuniene, Vytautas Magnus University
Domantas Jasilionis, Max-Planck Institute for Demographic Research and Vytautas Magnus University

Education is recognized as one of the key determinants of childbearing behavior, affecting both the timing and quantum of fertility. The existing evidence on educational differences in fertility provides contradictory results suggesting about U-shape and inverse gradients. These contradictory results depend on the type of fertility measures (cohort or period fertility) and can be attributable to both specifics and limitations of the data used for analyses as well as to a possible impact of tempo distortions on period fertility measures. Most of the existing analysis on the underlying fertility determinants has been performed on the basis of survey data, such as the FFS and the GGS. Alongside numerous advantages of survey-based evidence such as a large number of explanatory variables, there are several important disadvantages, including low response rates, low representativeness, and exclusion of some specific population groups. In addition, due to limitations of sample size, survey data often provide very limited possibilities to derive statistically robust fertility estimates by socio-demographic groups. Our study demonstrates potentials of census-linked fertility data for estimating robust and nationally representative parity-specific period and cohort fertility measures by education. Using a unique census-linked dataset (one of the first of this type in the Central and Eastern European region) covering entire population of Lithuania, the study provides new evidence and demographic insights into the scarce existing literature on educational differentials in period and cohort fertility in Lithuania. The paper also examines methodological issues related to estimation of period fertility measures by education as well as their impact on the observed period fertility differentials.