This study compares fertility rates from the Sample Registration System (SRS) and the 1992-93 National Family Health Survey (NFHS) in India as a means of determining the speed of fertility decline. Fertility since the 1970s declined faster based on the SRS than the NFHS. The present level of fertility is expected to be higher than recorded in either data source. Misreporting of women's ages in both sources warrants use of the general fertility rate (GFR), annual births divided by the estimated

more »

... id-year population of women aged 15-49 years. During 1978-92, GFRs agreed quite well for the period 1988-92, but in the preceding 5-year periods the ratio of the NFHS and the SRS increased from 1.00 in 1988-92 to 1.14 in 1984-87 to 1.19 in 1978-82. The GFR estimated from the NFHS was 10% higher than the rate estimated from the SRS. The underestimation in the SRS is attributed to underreporting of female births. Frequent omissions of female births drives the sex ratio higher than the average of 105-107 male per 100 female births. In fact, in 1997 the SRS sex ratio at birth for 1981-90 was 110 per 100. Evaluation studies of SRS data reveal an improvement in birth registration over time but did not examine the sex ratio over time. Sex ratios at birth from NFHS data were stable over time at around 106 per 100. Annual estimates of births from NFHS data show peaks and troughs that suggest misreporting of children's ages. Heaping occurs at ages 5, 8, 10, and 12 years. The increase in the GFR between 1983-87 and 1988-92 suggests omission of recent births or displacement of them in the earlier period. Couple protection rates (CPR) increased from 23% in 1978 to 44% in 1992. The total fertility rate (TFR) estimated from the statistical model tested on over 90 countries (TFR = 7.2931 - 0.0700 CPR) declines more steeply than the SRS rates. TFR from CPR estimates were not as steep as the NFHS rates.