Abstract: While the very low fertility rates obtained from the various demographic surveys in the 1990s including the 2000 census are considered to be beyond belief,the very high sex ratios at birth(SRB)reported from the same surveys are widely accepted.The major objective of this article is to re-estimate China’s SRBs in the 1990s using the educational data by examining the sex differences in birth under-reporting in the 2000 census.Results show that the estimated SRBs were not as high as those reported in the census.The actual SRBs and sex ratios at younger age groups are estimated to be lower than those reported in the census by 5-9 percentage points.Under-reporting of female births is estimated to account for at least 50% of the excessive amount of the SRBs in the 1990s.Socio-economic differentials in the SRB are also examined in this article using the micro data of the 2000 census.

Keywords: Sex ratio at birth, Census data, Educational data, Birth under-reporting

摘要： 当1990年代各种人口调查"过低的"生育水平难以置信时,其"过高的"出生性别比数据却被广泛接受。本文的主要目的是通过对2000年人口普查出生数据漏报的性别差异的分析,利用教育统计数据,重新估计2000年及90年代的出生性别比。结果表明我国实际的出生性别比并不像2000年普查数据反映的那样严重偏高。我国实际的出生性别比和低年龄组性别比要比普查反映的水平低5～9个百分点。1990年代在我国出生性别比偏高的部分中,女婴漏报的作用高达50%以上。本文还根据2000年普查千分之一抽样数据,考察了出生性别比的社会经济差异。

关键词: 出生性别比, 普查数据, 教育数据, 出生漏报