Measuring low fertility: rethinking demographic methods

Abstract

In a low fertility context demographic characteristics like the parity distribution of the population and the timing of births can lead to substantial changes in period fertility. In this article we review the literature on the analysis of fertility by parity and on tempo distortions. Both approaches can be combined through the use of the parity and age fertility model as recently proposed by Kohler and Ortega (2002a). We show applications of the approach for the analysis of both period and cohort fertility. For the analysis of period fertility a new measure is proposed, the Period Fertility Index, which is free from tempo distortions and from the influence of the population parity distribution. The Period Fertility Index is connected to the TFR by two ratios: the mean tempo effect, which generalizes previous work by Bongaarts and Feeney (1998), and the parity composition effect. Besides the period fertility index, all standard life table functions can be defined as synthetic cohort measures including mean ages at birth, interbirth intervals, parity progression rates or the stable parity distribution. From the perspective of cohort fertility it is possible to determine future childbearing intensities for conditional postponement scenarios. These can then be used for cohort completion or fertility forecasting. Examples of the different measures and techniques are given based on recent Swedish data.