China is aging fast, and for older households the principal micro-data source is the China Health and Retirement Longitudinal Study (中国健康与养老追踪调查). CHARLS follows a nationally representative sample of Chinese residents aged 45 and older (and their spouses) across 28 provinces, with baseline interviews in 2011 of roughly 10,000 households and 17,500 individuals, and follow-ups roughly every two to three years since. It is harmonized with the global family of aging studies anchored by the U.S. Health and Retirement Study (HRS) — including ELSA, SHARE, KLoSA, MHAS, IFLS, JSTAR, and SAGE — which makes China’s aging trajectory directly comparable with other countries’.
The most recently released round is Wave 5 (2020), announced on the official portal on 16 November 2023. As of September 2026, a Wave 6 covering 2023/2024 has not been confirmed as fielded or released.
At a glance
- Conducting institution: National School of Development, Peking University (formerly the China Center for Economic Research, CCER).
- Coverage period: 2011–present; waves roughly every 2–3 years (baseline 2011; 2013; 2015; 2018; 2020), plus a 2014 Life History Survey module.
- Latest released wave: Wave 5 (2020) — per the official portal’s news list, “CHARLS Wave 5 (2020) National Followup Survey Data has been Released,” 16 November 2023. Wave 6 (2023/2024): unconfirmed.
- Geographic coverage: national — 28 provinces in published descriptions; multi-stage PPS sampling across 150 county-level and 450 village-level units.
- Unit of observation: individual respondents aged 45+ and their spouses; households.
- Sample scope: roughly 10,000 households / 17,500 individuals at baseline across 150 counties/districts and 450 villages/resident committees (official portal figures). Response rates above 86% across waves (per published papers).
- Update frequency: every 2–3 years.
The data collected
Health status (self-reported health, activities of daily living / instrumental ADL, chronic disease, pain), physical measurements, cognition (an 11-test battery, MMSE-based), depression (CESD-10), medical insurance, healthcare utilization, employment and retirement, pensions, household income and assets, intergenerational transfers, social activities, and life-course history (the 2014 module covers education and work history).
Survey design and ethics
Face-to-face longitudinal interviewing with stratified multi-stage probability-proportionate-to-size sampling and computer-assisted interviewing (CAPI). Ethics approval by the Biomedical Ethics Review Board of Peking University (NO. IRB 00001052–11015), with written informed consent from participants (per a BMJ Open paper). Harmonized with the HRS family of aging studies via linked sister studies on the official portal.
Access
- Access point: the official portal, http://charls.pku.edu.cn/en — data downloads via charls.charlsdata.com; user forum at forum.charlsdata.com. A baseline user guide is publicly available (User2011.pdf on the portal). Link: http://charls.pku.edu.cn/en
- Access terms: registration/application-based — peer-reviewed papers describe the database as “free and open to society; download after application and review.” Not paywalled.
- Canonical citation: Zhao YH, Hu YS, Smith JP, Strauss J, Yang GH. Cohort profile: the China health and retirement longitudinal study (CHARLS). Int J Epidemiol. 2014;43:61–68. doi:10.1093/ije/dys203.
Research uses
The core micro-source for reporting on aging China, pension-system performance (Basic Pension, New Rural Social Pension), health equity and rural–urban health gaps, retirement and labor-force exit, disability, and long-term-care demand. It complements official NBS/MOHRSS statistics with panel micro-evidence. Because of the HRS harmonization, it is also the dataset that makes China-international comparisons of aging and pension outcomes possible.
Where it stops — scope limits
- It covers only adults aged 45+ and their spouses. Do not use it for general-population fertility, youth employment, or child development — use CFPS or CGSS for those.
- It is not a health-care utilization register: utilization is self-reported, and it does not directly measure health-system inputs such as hospital budgets or staffing, nor policy implementation.
- Respondents’ children and grandchildren appear only via transfers and living arrangements, not as units of observation.
- Death is not directly observed between waves; attrition/exit data are limited. Over 13+ years of follow-up the panel is unbalanced.
- The newest released wave (2020) predates the most recent policy years; with Wave 6 unconfirmed, recent-policy analysis cannot rely on post-2020 CHARLS data.
- Analytic sample sizes vary by module (e.g., the cognitive batteries are not complete for all respondents).
Known revisions
- Earlier English expansions sometimes wrote “China Longitudinal Study of Health and Retirement” (seen in a PMC article); the official canonical expansion is “China Health and Retirement Longitudinal Study.”
- Wave 6 (2023/2024) fielding/release: unconfirmed as of September 2026.
Authoritative documentation
- Official CHARLS portal (identity, conducting institution, wave-release news 2023-11-16, baseline design, data repository links): http://charls.pku.edu.cn/en
- Zhao et al. 2014 cohort profile (sampling design, baseline, citation): doi:10.1093/ije/dys203
- BMJ Open 2026 paper (sampling design, waves 2011/2013/2015/2018/2020, ethics); Frontiers in Public Health papers (variable domains)