Period Paramount? A Critique of the Cohort Approach to Fertility

The aim of this discussion of cohort and period fertility estimation was to clarify the issue and resolve the conflict between 2 approaches. The typical view of these approaches is that fluctuations in period fertility represented in the total fertility rate (TFR) are due to the timing and pace of cohort fertility; these TFR changes distort. TFR appears high when the pace of cohort fertility increases. This tempo component of fertility must be removed through adjustment in cohort and period measures. Intercohort change reflects the true time path of fertility. The conventional view has been comprehensively discussed by such persons as Hajnal and others who have pointed out that age specific fertility rates are more consistent in period than in cohort analysis. The weight of statistical evidence supports the notion that all birth and marriage cohorts react as a single unit to the factors that determine the general level of fertility at that time. The cohort approach continues to be used because it offers more options for interpretation complements period analysis and has stronger arguments for its use. The resilience in the use of the cohort approach was examined historically by focusing on the demographic role of the cohort perspective and identifying the parts of current thinking on the demography of the period/cohort issue. Ryder has concertedly argued that fertility is affected by past reproductive history or life cycle stage that cohorts are distinctive that theories of fertility use a cohort framework that time series of cohort TFRs are smoother that the cohort framework is a natural approach to behavior and that changes in cohort fertility reflect changes in period amount. Several other views are indicated but the 6 arguments are discussed individually at some length. The problems with period fertility are indicated and an alternative approach to fertility measurement is presented: the period parity progression approach not as an index for a period but as the schedule of duration specific probabilities. The data issues and the role for the cohort method are discussed. The conclusion is that the confusion that has arisen is 2-fold: confusion between time trends in fertility and time trends in reproductivity. Confusion has also arisen due to Hajnals solution to short-term fluctuations in period fertility.