Date

2019-01-01

Author

Kırdar, Murat Güray
Tunalı, Insan
Dayıoğlu Tayfur, Meltem

Metadata

Show full item record

Item Usage Stats

228
views

0
downloads

Cite This

We study the aggregate labor force participation behavior of women over a 25-year period in Turkey using a synthetic panel approach. In our decomposition of age, year, and cohort effects, we use three APC models that have received close scrutiny of the demography community. We rely on predictions from just-identified models that render different methods comparable. The exercise is carried out by rural/urban status and by education to tease out some key differences in behavior, and to test hypotheses about the course of participation. Our comparative methodology yields remarkably consistent profiles for most subsamples, but not all. Notably all methods reveal an M-shaped age profile attributable to child-bearing related interruptions in rural areas and for low-educated women in urban areas. We also find that younger cohorts among the least-educated women are more likely to participate, contrary to the belief that culture stands in the way. This implies that the recent rise in the aggregate participation rates is not only due to a composition effect arising from increasing education levels. We also show that Turkey has reached the turning point of the U-shaped pattern in female participation. In addition, we dwell on methodological issues and offer explanations for the fragility of the methods. We establish that evolution of the linear trend present in the crosssection age profiles is responsible for the apparent differences in the findings.

Subject Keywords

Female labor force participation, U-hypothesis, Synthetic birth cohort analysis, Age-participation profiles, Cohort effects, M-shaped profile, Culture

URI

https://hdl.handle.net/11511/88418
http://ftp.iza.org/dp12844.pdf

Collections

Department of Economics, Article