【作者】 李成；

【导师】 米红；

【作者基本信息】 浙江大学 ， 人口学， 2021， 博士

【摘要】 研究背景根据《World Population Prospects 2019》公布的数据,1950年至2015年中国人口死亡水平经历了快速降低,出生预期寿命平均每年增长0.5岁,是联合国对相同时期内对其他发展中国家平均预测水平的1.2倍。尽管如此,对中国及各省市死亡转变过程的研究却很少。分析这段时期内中国死亡转变过程,有助于深入了解中国社会经济发展过程中人口死亡率的动态变化。研究目的和内容本研究的目的:其一为获得较为准确的分省分性别生命表;其二为了解1981年至2015年中国及分省死亡转变情况;其三为评估三套常用的模型生命表和五种死亡率模型对中国的适用性。研究内容分为三个部分:一是使用半直接修正法、经验公式法和模型拟合法调整人口普查和抽样调查婴幼儿、成人和高龄死亡率并进一步编制1981年至2015年分省分性别生命表;二是从预期寿命、死亡模式、预期寿命变化中死亡过程以及死亡原因四个方面分析1981年至2015年中国和选定省市的死亡转变;三则是利用本研究编制的1981年至2015年中国及分省分性别生命表评估CoaleDemeny、United Nations、中国分区三套区域模型生命表以及Clark-Sharrow、ModLogit、Log-Quad、augment-Log-Quad以及SVD-Comp五种死亡率模型对中国的适用性。数据和方法本研究使用的数据分为三大类:一是1982后历次人口普查和1%人口抽样调查死亡数据;二是死因数据,其中包括Institute for Health Metrics and Evaluation的中国死因数据、1973-1975年全国第一次死因调查和中国死因监测数据集;三是China subnational Millennium Development Goal 4儿童死亡率数据,用以修正普查和1%抽样调查中的婴幼儿死亡率。除此以外所用到的人口和经济数据则来自统计年鉴和网络等公开渠道。根据死亡转变理论,分析方法采用死亡水平指标分解来确定中国及分省死亡转变的每个阶段,同时借助聚类分析对死亡转变过程中分省死亡模式的相似性进行讨论,并基于比例风险模型对选定省市人口不同时期的死亡率异质性进行分析。研究结论1.以1990年和2000年为节点,中国在1981年至2015年完成了“婴儿死亡率降低—成人死亡率降低—老龄死亡率降低”的完整死亡转变过程;2.整体各阶段的死亡转变实际是局部完整转变阶段叠加的结果;3.分省死亡模式可分为西南地区、西北地区、东北地区和东南地区,以及直辖市模式,而且这五种主要死亡模式随时间推移发生了一定程度的变化;4.所有省市整体所表现出的死亡转变特点与全国特点一致,但具体省市的死亡转变过程有很大差异。总体上东部地区省市的死亡转变进程要快于中西部地区省市;5.死亡率异质性对各省市人口作用较弱,但死亡转变过程中死亡率异质性发生了系统性改变;6.全国层面养老保障政策的重心应从老龄人口向高龄人口过渡,并且应以进一步提升老年人生活福祉为主。7.省级层面养老保障政策制定应在参照全国层面政策的基础上,根据各自情况选择不同的目标和主要保障对象。8.三套模型生命表对中国的适用性应具体研究内容和性别而异,且不同模型生命表死亡模式同实际情况的偏差具有一定的相似性。五种死亡率模型能够较好的预测中国人口45岁以上年龄别死亡率。从精度和稳定性角度考虑,优先选择augment-Log-Quad模型。研究贡献本研究的理论贡献在于对流行病转变理论、人口异质性理论和人口学统计学分解、聚类分析方法的整合,较为详细地阐释了中国人口死亡转变的规律,并为公共服务体系特别是社会保障体系中的养老、医疗和长期照护等提供一些人口理论上的参考。方法和结果上的贡献则在于采用横向研究与纵向研究相结合的方式对中国及分省死亡转变问题予以探讨,并将各种数理人口学分析技术根据研究理论和问题需要灵活使用;编制了较为准确的分省分性别生命表,为今后的大数据分析决策系统提供了数据参考;并为研究中国人口时模型生命表和死亡率模型的选择提供了一定的借鉴。更多还原

【Abstract】 Research BackgroundAccording to World Population Prospects 2019,the population in China has experienced the rapid mortality reduction between 1950 and 2015,with life expectancy at birth increasing by 0.5 years per year on average,1.2 times higher than the United Nation’s projection for other developing countries over the same period.Despite this,little research has been done on the mortality transition in China and its provinces.Analysis of the mortality transition in China during this period can provide insight into the dynamics of mortality during China’s socio-economic development.Research Objectives and ContentsThis study has three objectives: First,to acquire more accurate province-sexspecific life tables;Second,to understand mortality transition in China and its provinces from 1981 to 2015;Third,to assess the applicability of three model life tables and five mortality models to China.The contents include three parts: First,using the semi-direct correction,the empirical formula and the model fitting method to adjust child,adult and old-age mortality rates in the census and sample survey,and further constructing province-sexspecific life tables from 1981 to 2015;Second,analyzing the mortality transition in China and selected provinces from 1981 to 2015 in four aspects: life expectancy,mortality pattern,mortality dynamics in the change of life expectancy,and causes of death;Third,using the provincial sex-specific life tables from 1981 to 2015 developed in this study,the applicability of the Coale-Demeny,United Nations,and China Reginal model life tables and the five mortality models of Clark-Sharrow,Mod-Logit,LogQuad,augment-Log-Quad,and SVD-Comp to China was evaluated.Data and MethodsThe data used in this study divides into three major categories: First,mortality data from all previous censuses and 1% sample surveys after the year of 1982;Second,cause of death data,containing China’s cause of death data from the Institute for Health Metrics and Evaluation,the first national cause-of-death survey,1973-1975 and China’s Cause of Death Surveillance Dataset;Third,child mortality data from China subnational Millennium Development Goal 4,which is used to correct the infant and child mortality rates from the census and the 1% sample survey.Besides,other data like demographic and economic data was obtained from publicly available sources such as statistical yearbooks and the Internet.According to the mortality transition theory,the analysis is based on the decomposition of mortality indicators to identify each stage of mortality transition in China and its provinces,meanwhile the similarity of mortality patterns between provinces during the mortality transition was discussed by the clustering analysis,and the mortality heterogeneity of selected provincial populations over time was analyzed through proportional risk models.Research Conclusions1.China completed a complete mortality transition from 1981 to 2015 with the process of “infant mortality decline-adult mortality decline-old ages mortality decline”,taking 1990 and 2000 as the nodes.2.The whole mortality transition is actually the result of the accumulation of local complete transition stages.3.Provincial mortality patterns can be divided into southwest,northwest,northeast,and southeast,as well as municipalities patterns,and these five major mortality patterns have changed to some extent over time.4.The overall mortality transition characteristics of all provinces are consistent with those of the nation,but the detailed mortality transitions of selected provinces vary greatly.In general,the mortality transition is faster in the eastern provinces than in the central and western provinces.5.The effect of mortality heterogeneity on China’s provincial population is nonsignificant,but mortality heterogeneity has changed systematically during the mortality transition.6.The focus of old-age security system at the national level should shift from the older to the more senior population and should be aimed at further enhancing the wellbeing of older people.7.The formulation of provincial pension system should refer to the national policies,choosing different objectives and main targets for security according to their situations.8.The applicability of the three model life tables to China varies with research aims and sex,and the deviations of mortality patterns from the actual situation are somewhat similar across the model life tables.The five mortality models are able to accurately predict the mortality rate of the Chinese population over the age of 45 years.From the perspective of precision and stability,the augment-Log-Quad model is preferred.Research ContributionsThe theoretical contribution of this study lies in the integration of epidemiological transition and population heterogeneity theories,and demographic decomposition,cluster analysis to explain in some detail the process of mortality transition in China,and the provision of some demographic theoretical references for the public service system,especially for old age,medical care and long-term care in the social security system.The contribution of the methodology and results lies in the combination of crosssectional and longitudinal analysis to explore the mortality transition in China and its provinces,and the flexible use of various mathematical demographic techniques based on the needs of the research theories and problems,as well as the construction of relative accurate province-sex-specific life tables,which provides a data source for big data analysis and decision-making systems in the future.And this thesis also provides a reference for the selection of model life table and mortality models when studying China’s population.更多还原