Denominator adjusted PGE fertility and mortality rates for Pakistan and Bangladesh: 1962-65
Afzal, M.
Pakistan Development Review 16(2): 192-207
1977
ISSN/ISBN: 0030-9729 PMID: 12260625 DOI: 10.2307/41258369Document Number: 205348
The Population Growth Estimation (PGE) experiment yielded 3 series of fertility and mortality rate estimates for Pakistan and Bangladesh for 1962-1965: 1) the series based on longitudinal registration (LR); 2) the estimates based on the cross-sectional survey (CS); and 3) the Chandra-Deming (CD) estimates obtained through the matching of LR and CS. The CD estimates were inflated because, while the numerator was adjusted for underregistration and underenumeration, the denominator remained unadjusted. Almost constant relative differences of about 20% were found between the CD and LR estimates. To work out the denominator-adjusted PGE birth and death rates, the present study used the ratio of births to deaths, not their absolute numbers. The resulting denominator-adjusted CD crude birth rates for both Pakistan and Bangladesh were 46, compared with the unadjusted rates of 52 and 53, respectively. Similar adjustments in LR rates have yielded vital rates of almost the same order of the CD rates, supporting the validity of this approach. The rates obtained provide a more realistic basis for preparing population projections. Incorporation of similar adjustments in the age-specific mortality rates can lead to a revised set of life tables, and the age distribution of the 1972 census can be used with the estimates of fertility and mortality obtained in the present study to project vital rates for the 2 countries. In addition, the adjusted estimates of the crude birth rate and age-specific fertility rate provide more realistic benchmarks for evaluating the impact of Pakistan's family planning program. Future PGE-type projects can provide realistic fertility and mortality estimates if 1) the registration and enumeration of vital events are accurate enough to yield the correct ratios of the number of births and deaths; and 2) the population counts in at least the initial and terminal years of the experiment are of the same degree of accuracy, so as to yield a correct average rate of population growth.