The study of China's fertility has been in a strange situation since the early 1990s. On the one hand, the growing number of censuses, fertility surveys and annual surveys of population change should provide sufficient data for detailed investigation of recent fertility decline; on the other hand, key fertility data are increasingly affected by problems of under-reporting and internal inconsistency that form an obstacle to such investigations. This is partly due to the fact that while the statistical authority has published some adjusted fertility statistics, it has been rather reluctant to release details of the technique used to make such adjustments since the late 1990s. As a consequence, researchers sometimes have to rely on limited information and various assumptions to estimate fertility, which has inevitably led to disparate results (Cai, 2008; Goodkind, 2004; Guo, 2004; Guo and Chen, 2007; Retherford et al., 2005; Scharping, 2007; Wang, 2003; Yu and Xie, 2000; Zhang, 2004; Zhai and Chen, 2007; Zhang and Cui, 2003; Zhang and Zhao, 2006). While some of these efforts have provided insights into China's recent fertility trends, this unsatisfactory situation and the controversy arising from it have created confusion, especially among users who rely primarily on secondary data sources for their information on China's fertility level. By comparing various officially reported fertility rates and reconstructing China's recent fertility statistics, this article further investigates the problems of collecting and publishing fertility data encountered since the early 1990s.(1) It then examines China's recent fertility level and addresses a number of related issues.
I. Major problems in China's recent fertility statistics
We will begin by looking at the problems associated with China's recent fertility statistics and the controversies surrounding China's recent fertility level.
First, China's statistical authority has routinely released the number of births and the crude birth rate (CBR) in recent years. The total fertility rate (TFR) has also been reported in official publications, though in a less regular fashion. These fertility statistics have often been adjusted for under-reporting or estimated on the basis of results obtained from the annual survey of population change (also referred to as annual population change survey or annual survey). However, details on how such adjustments or estimations are made have not been available since the late 1990s, (Zhang and Zhao, 2006) so it is not easy to determine the accuracy of these statistics.
Second, the Chinese authorities publish the age-specific fertility rate (ASFR) as a part of the annual population change survey results.(2) These statistics are generally not adjusted for under-reporting. Because the information on underreporting was not released in most of the years, it is difficult to determine the extent to which these unadjusted statistics under-represent actual fertility. Since officially released numbers of births, CBRs and TFRs have been adjusted under certain assumptions while ASFRs have not, these fertility statistics, as expected, are not internally consistent. Such inconsistencies cannot be resolved easily, even if the proportion of unreported births in the total population is known, since under-reporting may not be distributed evenly among women of different ages. Different assumptions about the age patterns of under-reporting will result in different ASFRs and TFRs.
Third, the published findings of the annual population change survey also include the general fertility rate (GFR). If the GFR has not been reported directly, it can be computed easily from the number of births and number of women of reproductive ages reported by the annual population change surveys. These statistics are generally not adjusted for under-reporting. The GFR for the national population can also be computed from the adjusted number of births and the total number of women of reproductive ages, which would result in adjusted GFRs. …