China's Rural Population Statistics at the Local Level

Abstract

Written for those who use Chinese population data and want a better understanding of their provenance and reliability and those who may directly utilize local level materials in studies of Chinese population and social structure, this report describes the statistical system of 1 rural county, Shifang Xian in the Chengdu Plain of Sichuan Province. It is based on interviews with local government officials, on examinations of population records and reports at different levels of administration, and on a sample survey of households conducted in the winter and spring of 1981. Until the mid 1970s, the primary source of China's rural population data was the household and vital events registers established in the 1950s. Following the formation of the Birth Planning Office in 1971, a separate reporting system of population statistics began to develop alongside the household registration system. The birth planning system uses the reports of team, brigade, and commune level cadres concerned with health work and women's affairs to provide a richer and more current set of vital events and birth planning data than the household registration system could provide. Discrepancies in data emerging from the 3 sources are bound to occur because of error and because of the different methods used in the compilation of data. Currently, there are 2 basic sources for population data in Shifang Xian: the monthly reports of the brigade level birth planning workers and the year end reports of the team accountants. The household and vital events registers, once central to population statistics, retain their legal role but have diminished importance for statistical purposes. There continues to be important questions about the operation of the statistical system. With increasing reliance on the newly developed birth planning statistical system for information on vital rates, Shifang has apparently moved from statistics based on date of registration derived from a de jure system of vital events registration to statistics based on date of occurrence derived from de facto or quasi de facto records of health workers and team accountants. As the latter system is more accurate, it seems likely that transition has been marked by discontinuities in time series of population counts and vital rates. The statistical system observed in Shifang is administered by highly motivated cadres. It apparently produces statistics of good quality. Linkage to the economy, constant updating, surveillance of pregnancy, and a level of overlap give the overall system considerable strength and should not produce substantial errors under normal circumstances. A serious remaining problem, i.e., the underreporting of infant deaths, is primarily due to inadequate training of statistical workers. Shifang Xian is an unusual county, and no conclusions should be drawn about the accuracy of rural population statistics based on performance there.