Regression estimates of the gross reproduction rate using moments of the female age distribution
Gunasekaran, S.; Palmore Ja
Asian and Pacific Census Forum 10(4): 5-10
1984
ISSN/ISBN: 0732-0515 PMID: 12339581 Document Number: 416920
The 2 most common regression methods for estimating fertility are those developed by Rele (1967) and Bogue and Palmore (1964) as revised by Palmore (1978); these methods are sometimes more useful than others because they employ the data that most censuses collect--especially the child-woman ratio by subnational geographic areas. Both the Rele and Palmore equations base their estimates on the child-woman ratio (CWR), the ratio of children ages 0-4 to women ages 15-49. However, in countries where children ages 0-4 are severely undercounted, these estimates will reflect the errors. Furthermore, the Palmore equations cannot be applied if information on marital status and the infant mortality rate is missing, or will reflect the errors if that information is inaccurate. India and many other countries have these data problems. The authors therefore present an alternate method for estimating the gross reproduction rate from the age distribution of females and an estimate of female life expectancy at birth, and apply this new method to Indian state data for 1961 and 1971. Finally, the results are compared to the other estimates by Rele, Bogue-Palmore and Premi (1974) for Indian states. Findings suggest that the authors' Model 6 may be a useful method for estimating fertility if age data for ages 0-4 are suspected to be inaccurate and if data are available for estimating female life expectancy at birth. No assumptions are made about the stability of the population, but it may work better when fertility has not fluctuated greatly.