| 355 | 0 | 605 |
| 下载次数 | 被引频次 | 阅读次数 |
人口负增长、少子化和老龄化给我国人口安全、经济发展和社会保障等各方面带来了挑战和风险。本文在传统队列要素法高、中、低方案确定性参数设定的基础上，基于“充分覆盖”的想法，提出一种针对总和生育率不确定性的稳健估计方法，并结合队列要素法，在高、中、低三种总和生育率水平下，对我国2026—2100年人口趋势进行预测并给出置信区间。研究发现，未来我国人口总量将持续下降，少子化和老龄化程度将进一步加深。短期而言（2026—2050年），三种方案预测差异较小，表明总和生育率的变化对短期人口趋势的影响有限；长期而言（2051—2100年），三种方案预测显著分化，低方案形势更为严峻，而高方案较中方案改善明显，表明总和生育率的变化对中长期人口形势具有重要影响。敏感性分析表明，总和生育率的小幅提升能明显改善长期人口形势；生育激励政策具有明显的“窗口效应”，实施越晚，政策效果损失越大；受育龄妇女队列结构的变化，生育年龄的推迟在中长期反而能小幅缓解人口下降和老龄化进程；预期寿命的增长虽然提高了中长期人口规模，但增量主要来自老年人口扩张。据此，本文提出完善生育支持政策，抓住2026—2035年政策窗口期，加强人口动态监测预测的政策建议。本文研究可为积极应对人口负增长、少子化和老龄化挑战，促进人口长期均衡发展和经济社会可持续发展提供数据基础和决策参考。
Abstract：Population decline, low fertility, and population aging pose challenges and risks to China's demographic security, economic development, and social security systems. Building on the conventional cohort-component method with deterministic parameter settings under high, medium, and low scenarios, this paper proposes a robust estimation approach to quantify uncertainty in the total fertility rate(TFR), guided by the “full coverage” principle. Using the cohort-component method, this paper projects China's population for 2026—2100 under three TFR scenarios—high, medium, and low, and constructs corresponding confidence intervals. The results indicate that total population will continue to decline with further declines in fertility and intensifying population aging. In the short term(2026—2050), differences across the three scenarios are small, implying that changes in the TFR have limited influence on short-term population trends. While in the long term(2051—2100), the projections diverge markedly: the low-TFR scenario yielding a much more severe outcome, whereas the high-TFR scenario delivers substantial improvements compared to the medium-TFR scenario, indicating that TFR shifts materially shape the medium-to long-term demographic trajectory. Sensitivity analysis shows that even small increases in the TFR can substantially improve the long-run demographic outcome; pronatalist policies exhibit a clear “window effect”: the later they are implemented, the greater the loss in policy effectiveness; moreover, driven by changes in the cohort structure of women of childbearing age, postponing childbearing may slightly mitigate population decline and aging in the medium to long term; although rising life expectancy increases population size over the medium and long term, the gains are concentrated mainly among older age groups. Finally, this paper recommends further improving fertility support policies, seizing the policy window from 2026 to 2035, and strengthening dynamic population monitoring and forecasting. This study provides a data foundation and policy reference for actively addressing the challenges of population decline, low fertility and population aging, and for promoting long-term demographic balance and sustainable socioeconomic development.
[1]贺丹.中国人口展望（2024）:迈向高质量发展的未来[M].福州:福建人民出版社, 2024.
[2]张现苓,翟振武,陶涛.中国人口负增长:现状、未来与特征[J].人口研究, 2020, 44(3):3–20.
[3]Alkema L, Raftery A E, Gerland P, et al. Probabilistic Projections of the Total Fertility Rate for All Countries[J]. Demography, 2011, 48(3):815–839.
[4]Azose J J, Raftery A E. Bayesian Probabilistic Projection of International Migration[J]. Demography, 2015, 52(5):1627–1650.
[5]Bai R, Liu Y, Zhang L, et al. Projections of Future Life Expectancy in China up to 2035:A Modelling Study[J]. The Lancet Public Health, 2023,8(12):e915–e922.
[6]Bohk-Ewald C, Li P, MyrskyläM. Forecast Accuracy Hardly Improves with Method Complexity when Completing Cohort Fertility[J].Proceedings of the National Academy of Sciences, 2018, 115(37):9187–9192.
[7]Booth H. Demographic Forecasting:1980 to 2005 in Review[J]. International Journal of Forecasting, 2006, 22(3):547–581.
[8]Booth H, Pennec S, Hyndman R. Stochastic Population Forecasting Using Functional Data Methods:The Case of France[C]. Annual Meeting of the International Union for the Scientific Study of Population, Marrakech, Morocco, 2009.
[9]Hyndman R J, Booth H. Stochastic Population Forecasts Using Functional Data Models for Mortality, Fertility and Migration[J]. International Journal of Forecasting, 2008, 24(3):323–342.
[10]Lee R D. Modeling and Forecasting the Time Series of US Fertility:Age Distribution, Range, and Ultimate Level[J]. International Journal of Forecasting, 1993, 9(2):187–202.
[11]Lee R D, Tuljapurkar S. Stochastic Population Forecasts for the United States:Beyond High, Medium, and Low[J]. Journal of the American Statistical Association, 1994, 89(428):1175–1189.
[12]Leslie P H. On the Use of Matrices in Certain Population Mathematics[J]. Biometrika, 1945, 33(3):183–212.
[13]Preston S, Heuveline P, Guillot M. Demography:Measuring and Modeling Population Processes[M]. Oxford:Blackwell, 2001.
[14]Raftery A E, Alkema L, Gerland P. Bayesian Population Projections for the United Nations[J]. Statistical Science, 2014a, 29(1):58–68.
[15]Raftery A E, Lalic N, Gerland P. Joint Probabilistic Projection of Female and Male Life Expectancy[J]. Demographic Research, 2014b, 30:795–822.
[16]Raftery A E, Li N,ŠevčíkováH, et al. Bayesian Probabilistic Population Projections for All Countries[J]. Proceedings of the National Academy of Sciences, 2012, 109(35):13915–13921.
[17]ŠevčíkováH, Li N, KantorováV, et al. Age-Specific Mortality and Fertility Rates for Probabilistic Population Projections[M]//Schoen R, ed.Dynamic Demographic Analysis. Cham:Springer International Publishing, 2016:285–310.
[18]United Nations Department of Economic and Social Affairs, Population Division. World Population Prospects 2024:Methodology of the United Nations Population Estimates and Projections[R]. New York:United Nations, 2024.
[19]United States Census Bureau. Methodology, Assumptions, and Inputs for the 2023 National Population Projections[R]. Washington, D.C.:United States Census Bureau, 2023.
[20]Whelpton P K. An Empirical Method of Calculating Future Population[J]. Journal of the American Statistical Association, 1936, 31(195):457–473.
(1)人口更替水平总和生育率是指维持人口代际更替所需的生育水平，通常约为2.1。
(2)按照国际通用标准，65岁及以上人口占比超过14%，即进入中度老龄化社会。相关网址为https://www.stats.gov.cn/zs/tjws/tjbz/202301/t20230101＿1903949.html。
(1)基准方法假设未来总和生育率与过去持平。
(2)人类死亡数据库（Human Mortality Database,HMD）由德国Max Planck人口研究所、美国加州大学伯克利分校和法国人口研究所共同建设和维护，提供各个国家的死亡率数据。相关网址为https://www.mortality.org。
(3)相关网址为https://www.un.org/development/desa/pd/data/model-life-tables。西区模式生命表是我国人口预测研究中的常用设定（张现苓等，2020；贺丹，2024）。
(1)因篇幅所限，队列要素法的计算过程以附录1展示，见《统计研究》网站所列附件。下同。
(2)净迁移人口为迁入人口减去迁出人口。
(1)因篇幅所限，模型检验以附录2展示。
(2)因篇幅所限，不同时间窗口下方法（B）对参数ρ的估计结果以附录3展示。
(3)根据2025年出生人口792万，推算出2025年总和生育率为0.99。
(1)因篇幅所限，不同时间窗口下对参数*的估计结果以附录4展示。
(1)因篇幅所限，不同预测起点，本文方法对总和生育率的预测结果和联合国预测结果的对比以附录5展示。
(2)2020年，我国部分城市总和生育率为上海0.73、北京0.87、天津0.92、成都1.05、深圳1.06。韩国2020年总和生育率0.84。2023年韩国总和生育率0.72,2023年上海市户籍人口总和生育率0.60。因篇幅所限，我国及其部分城市总和生育率和韩国总和生育率趋势以附录6展示。
(3)2000—2015年总和生育率的平均值约为1.31。
(1)因篇幅所限，双逻辑函数的参数估计、拟合效果、预测结果和残差诊断以附录7展示。
(2)因篇幅所限，插值算法以附录8展示。
(3)因篇幅所限，各年度人口迁移数据以附录9展示。
(1)2025年总人口和各年龄段人口数据来源于《2025年全国1%人口抽样调查主要数据公报》，与2026年初公布值略有差异。
(2)因篇幅所限，2075年和2100年预测结果以附录10展示。
(1)因篇幅所限，总和生育率长期均值水平提升对人口总量和结构的改善效果以附录11展示。
(2)人口总量型指标，是指反映特定人群规模的指标，如育龄妇女总量、劳动人口总量。
(1)人口结构型指标，是指反映人口结构的指标，如老龄化率、老年人口抚养比、年龄中位数、劳动人口年龄中位数。
(1)因篇幅所限，我国历史PASFR趋势以附录12展示。
(2)因篇幅所限，PASFR的预测模型构建、计算过程及预测结果以附录13展示。
(3)因篇幅所限，MAC延迟情景的人口预测结果以附录14展示。
(4)因篇幅所限，死亡率固定情景下，3种方案的人口预测结果以附录15展示。
基本信息:
DOI：10.19343/j.cnki.11-1302/c.2026.08.003
中图分类号:C924.2
引用信息:
[1]陈磊,陈松蹊,何婧.我国人口发展趋势及政策启示[J].统计研究,2026,43(08):31-45.DOI:10.19343/j.cnki.11-1302/c.2026.08.003.
基金信息:
国家统计局重大统计专项“融合抽样调查与多源大数据的人口统计理论和应用”（2025ZX11）; 国家自然科学基金重大项目课题“大规模商务场景下的数据科学理论”（72495122）;国家自然科学基金面上项目“超高维（条件）独立性检验及其在经济学中的应用”（72473114）; 教育部基础学科和交叉学科突破计划“海洋热–碳容量协同演变下的气候–生态系统临界点识别与可持续应对”（JYB2025XDXM801）
2026-08-25
2026-08-25