摘要
高质量人口死亡数据的匮乏长期困扰着中国的死亡率研究,死亡率的间接估计也由此成为中国人口研究的一项重要的基础性工作。过去常用的方法如模型生命表和死亡关联模型等存在的一些缺陷会导致死亡率的误估。Wilmoth等提出的二维死亡模型具有对数据要求少、估计效果好的特点,但该模型的提出、构建和验证均基于发达国家的死亡率数据,对中国人口死亡模式的适用性还有待检验。文章从模型构建、经验系数和输入参数3个方面考察该模型的适用性。研究结果表明,二维死亡模型在理论和模型估计上均对中国有较好的适用性,能准确地刻画中国人口死亡率的年龄模式。对拟合效果的检验和比较显示,二维死亡模型估计的精度优于其他的间接估计方法,拟合效果更好。
Lacking high quality data has been hindered mortality study in China for a long time, and hence mortality indirect estimation has become a fundamental work for morality study. Previously often used methods such as modal life tables and rational models of mortality are found to have certain flaws that cause the errors of the mortality estimation. A flexible two-dimensional mortality model has higher accuracy but requires less data than other methods. However, this model is constructed mainly based on the historical data of mortality in developed countries, it therefore needs to be inspected when applied to developing countries such as China, which lack good quality data. This paper investigated the application of the model in mortality estimation in China from three aspects, i.e., formulation of mathematical relations, sensitivity to empirical data and flexibility of entry parameters. The result shows that the two-dimensional mortality model performs very well in terms of flexibility and accuracy in estimating China's provincial mortality in 1981. In particular, using the coefficients derived from reliable provincial mortality data collected in the 1982 China population census, the model,even with one entry parameter, can yield very good fit. Our findings suggest that the two-dimensional mortality can be an effective tool for mortality estimation in China.
出处
《中国人口科学》
CSSCI
北大核心
2017年第1期81-91,共11页
Chinese Journal of Population Science
基金
国家自然科学基金面上项目"中国人口寿命不均等的变化趋势
影响因素及对策研究"(编号:71473044)的阶段性成果
作者简介
复旦大学人口研究所，副教授
硕士研究生
硕士研究生
