Various fertility measures in the recovery phase of total fertility rates in Japan

Miho Iwasawa, National Institute of Population and Social Security Research, Japan
Ryuichi Kaneko, National Institute of Population and Social Security Research, Japan

Japanese period total fertility rates (TFR), which had declined for over 30 years since the mid-1970s, shows explicit upturn after 2005 and it is largely due to increase in fertility rate in the late twenties and the thirties. In the presence of delayed childbearing, recovery of fertility rates, especially among the higher ages, should be explained by three possible factors. The first one is the inflation by the increase in the risk population at such ages caused by postponement of childbirth in the past time (structural effect/timing effect). The second one is the temporal effect reflecting short-time fluctuations in the socio-economic conditions. The last one is “genuine” change in fertility behavior which may affect cohort fertility level. To explain this TFR recovery, we show various fertility and nuptiality measures in both period and cohort perspectives and discuss their trends and the meaning of the differences between each measurement. As for period measures, in addition to conventional period TFR, we show fertility measures based on hazard rates of order-specific births. In this measure, the risk population is the person-years of women who have not experienced the event. Fertility measures based on hazard rates allow us to eliminate the amount attributable to the increase in the risk population due to delayed age schedule. Since temporal effect from socio-economic conditions should be independent from the change in the risk population, it would be more reasonable to use hazard-based measures when we examine temporal correlations with socio-economic conditions. Actually, to distinguish genuine change from temporal effect is both conceptually and practically difficult. We estimate age-specific fertility rates for cohorts in the process of reproduction using cohort age schedule model and use them as an auxiliary measures to identify temporal effects.