📄 medicine

Employment-type inequalities in healthy working life expectancy in China: evidence from CHARLS, 2011–2018

Using data from the China Health and Retirement Longitudinal Study (2011–2018), this study reveals that employment-type inequalities significantly shape healthy working life expectancy among older Chinese adults, with agricultural workers facing the highest burden of unhealthy work, particularly women, suggesting that retirement reforms must account for these disparities rather than relying solely on average longevity.

Original authors: Yulong Li¹, Panpan Wang¹, Qianyu Zhou², Mingyang Zhao², Ruoyao Huang³, Jiajun Chen, Hexin Zhang¹, Yang Yang¹, Yuxin Chen¹, Changqing Sun

Published 2026-10-19

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Yulong Li¹, Panpan Wang¹, Qianyu Zhou², Mingyang Zhao², Ruoyao Huang³, Jiajun Chen, Hexin Zhang¹, Yang Yang¹, Yuxin Chen¹, Changqing Sun

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

As populations age, governments around the world are looking at ways to keep people in the workforce longer. The logic is straightforward: if people live longer, they should work longer to support their own retirement and the pension systems that rely on them. But a longer life does not automatically mean a longer life in good health. A person might live to eighty, but if they spend the last fifteen years in pain or with limited ability to move, those extra years are not necessarily a benefit to the individual or the economy. This is where the concept of "healthy working life expectancy" becomes crucial. It is a measure that asks not just how many years a person will work, but how many of those years they will spend working while still in good health. The question for policymakers is no longer just whether older adults can work, but whether they can do so without suffering.

In China, this question is particularly urgent. The country is currently rolling out a gradual increase in the official retirement age, a move designed to address demographic shifts and pension pressures. However, the Chinese labor market is not uniform. It is divided into distinct groups: those who work for wages in formal jobs, those who run their own small businesses or family enterprises, and a vast number of people who work in agriculture. These groups face very different physical demands, have different levels of income security, and possess different abilities to stop working when they feel tired or unwell. To understand how the new retirement rules will affect real people, researchers needed to look beyond simple averages and examine how these different types of work shape the later years of life.

A team of researchers from Zhengzhou University and the First Affiliated Hospital of Zhengzhou University set out to map these differences using data from the China Health and Retirement Longitudinal Study. They analyzed information from over 11,000 adults aged 50 and older, collected across four different years between 2011 and 2018. Instead of simply asking who is working and who is retired, they created a more detailed picture by combining two factors: a person's current work status and their physical health. They defined "healthy" as having no limitations in daily activities like bathing or dressing, and "unhealthy" as having some difficulty with these tasks. By crossing these two factors, they identified four possible states for any older adult: working while healthy, working while unhealthy, not working while healthy, and not working while unhealthy.

The researchers then used these observations to estimate how many years a person could expect to spend in each of these four states from age 50 until the end of their life. To ensure their numbers were accurate and not skewed by missing data, they anchored their calculations to official national life tables, which provide a reliable baseline for how long people in China typically live. This allowed them to see the full distribution of a person's remaining years, broken down by whether they were working, whether they were healthy, and what kind of job they held.

The results revealed a stark divide that depends heavily on both gender and the type of work performed. For men, the differences were noticeable but less extreme. Men who worked as employees or in other wage jobs could expect to spend about 17.1 years working in good health, followed by about 1.4 years working while unhealthy. In contrast, men working in agriculture could expect to spend nearly 19 years working in good health, but they also faced a significantly higher burden of working while unhealthy, with an expectation of 4.0 years spent working despite physical limitations.

For women, the inequalities were far more pronounced. Women working in agriculture faced the most difficult profile of all. They could expect to spend only about 16.1 years working in good health, while simultaneously facing the highest expectation of working while unhealthy: nearly 7.7 years. This means that for a significant portion of their later lives, these women are likely to be working while struggling with physical limitations. On the other end of the spectrum, women who were self-employed or ran family businesses had the most favorable outlook. They could expect to spend over 20 years working in good health and only about 3.2 years working while unhealthy. Women who were employees fell somewhere in between, with about 17.6 years of healthy work and 5.1 years of unhealthy work.

The study also ran a simulation to see how these findings might change the "ideal" retirement age for different groups. They created a scenario that valued years of good health highly, subtracted years of working while unhealthy, and added a small value for the act of working itself to represent income and social connection. Under this model, the age that maximized overall well-being varied wildly depending on the job. For men in agriculture, the optimal retirement age came out to 68. For men in self-employment, it was 74. For women in agriculture, however, the optimal age was just 55. For women in other jobs, it ranged from 68 to 72. This suggests that a single, uniform retirement age for the entire country would benefit some groups while potentially harming others, particularly rural women who would be forced to work many more years in poor health.

The authors emphasize that these findings do not mean retirement reform should be abandoned, but rather that it must be evaluated with a much sharper focus on equity. Relying solely on average life expectancy is insufficient because it hides the reality that some people are working longer not by choice, but because they lack the financial security or physical ability to stop. The study suggests that if the goal is to extend working lives, the policy must be accompanied by protections for those in physically demanding or informal jobs. This could include flexible retirement options for those with health issues, better income support for older agricultural workers, and opportunities to transition to lighter work. Without these measures, extending the retirement age risks turning the later years of life for the most vulnerable workers into a prolonged period of struggling through work rather than a time of healthy participation.

Technical Summary: Employment-type inequalities in healthy working life expectancy in China

Problem Statement
As China implements a gradual increase in the statutory retirement age (effective January 2025), a critical public health question arises: do the groups expected to work longer possess comparable health capacity, working conditions, and social protection? While extending working lives is a response to demographic aging and pension sustainability pressures, global evidence suggests that gains in total life expectancy are not always matched by gains in healthy life expectancy. Existing literature has established inequalities in Healthy Working Life Expectancy (HWLE) by sex, socioeconomic status, and region in China. However, there is limited evidence on how broad employment types—specifically the distinctions between agricultural work, self-employment/family-business work, and wage employment—shape the balance between working in good health versus working despite functional limitations. The current policy debate often relies on average longevity, potentially obscuring the unequal burden of "unhealthy work" (working with functional limitations) across different labor market sectors.

Methodology
This study employed a life-table-anchored prevalence-based strategy using pooled data from the 2011, 2013, 2015, and 2018 waves of the China Health and Retirement Longitudinal Study (CHARLS).

Sample: The analytic sample consisted of 34,644 person-wave observations from 11,626 adults aged 50–100 years.
State Definitions: The authors defined four mutually exclusive joint health-work states based on functional capacity (no limitations in Activities of Daily Living [ADL] or Instrumental Activities of Daily Living [IADL]) and current work status:
1. Healthy-Working (HW)
2. Healthy-Not Working (HN)
3. Unhealthy-Working (UW)
4. Unhealthy-Not Working (UN)
Employment Classification: Respondents were stratified into three employment types: Agricultural workers, Self-employed/Family-business workers, and Employees/Other workers.
Estimation Approach: Unlike transition-based multistate models, this study anchored survival to sex-specific United Nations World Population Prospects life tables (2022 revision) to avoid implausible estimates arising from incomplete mortality follow-up in survey data. Age-specific state prevalences were combined with life-table person-years to estimate the expected years spent in each state from age 50 onward.
Policy Simulation: An illustrative deterministic scenario was constructed to evaluate welfare-maximizing retirement ages. A value function ( $Welfare_\theta$ ) was applied: $E(HW + HN) - E(UW) + \theta \times E(HW + UW)$ . This function rewards healthy years, subtracts unhealthy working years, and assigns a positive value ( $\theta$ ) to work itself. Sensitivity analyses were conducted with $\theta$ ranging from 0.00 to 0.30.

Key Results

Life Expectancy Baseline: Remaining life expectancy at age 50 was fixed at 27.11 years for men and 31.76 years for women.
Inequalities by Employment Type:
- Men: Agricultural workers had the highest Healthy-Working Life Expectancy (HWLE) at 18.91 years but also the highest Unhealthy-Working (UW) expectancy at 4.00 years. Employees had the lowest HWLE (17.12 years) and the lowest UW (1.42 years).
- Women: The most pronounced inequality was observed among women. Agricultural women had the least favorable profile, with the lowest HWLE (16.06 years) and the highest UW (7.65 years). Conversely, self-employed/family-business women had the highest HWLE (20.78 years) and the lowest UW (3.22 years).
Welfare-Maximizing Retirement Ages: Under the primary value specification ( $\theta = 0.20$ $θ = 0.20$ ), the optimal retirement age varied significantly:
- Men: 68 years (Agricultural), 71 years (Employees), 74 years (Self-employed).
- Women: 55 years (Agricultural), 72 years (Employees), 68 years (Self-employed).
- When pooled, the optimal age was 70 for men and 58 for women.
Sensitivity: The optimal retirement age was highly sensitive to the value assigned to work ( $\theta$ ). As $\theta$ increased, the optimal age for women rose from 50 to 65, while for men it rose from 50 to 75.

Key Contributions

Differentiation of Work Quality: The study moves beyond binary work/non-work metrics by separating "healthy work" from "unhealthy work." It demonstrates that extended working life can represent either healthy participation or prolonged exposure to functional limitations, with the latter being particularly burdensome for specific groups.
Employment-Type Perspective: It adds a crucial dimension to Chinese HWLE research by showing that broad employment types (agricultural vs. self-employed vs. wage) capture significant heterogeneity in work organization, physical demand, and exit pathways that standard socioeconomic indicators may miss.
Methodological Robustness: By anchoring survival to external national life tables rather than survey follow-up, the study provides a transparent period measure of health-work distribution that avoids biases from incomplete death observation.

Significance and Claims
The paper claims that employment type is a major source of inequality in later-life health-work trajectories in China. The authors argue that evaluating retirement-age reform solely on average longevity is insufficient because it fails to distinguish between longer healthy participation and longer work despite functional limitations.

The findings suggest that a uniform national retirement age increase may have unequal consequences. Specifically, the study highlights that women in agricultural employment face the most disadvantaged profile, characterized by low healthy-working years and high unhealthy-working years. The authors conclude that a public health approach to retirement reform must account for the unequal burden of unhealthy work. They propose that equity-oriented reform should not only adjust statutory ages but also include protective measures, such as flexible retirement pathways, stronger income protection for informal workers, and workplace adaptations, to ensure that extended working lives are achieved through healthy participation rather than economic necessity.

Limitations Acknowledged
The authors note that the study provides a period summary rather than a cohort forecast and does not estimate causal effects of employment on health. The analysis relies on broad employment categories and functional health measures (ADL/IADL) which may not capture all dimensions of work ability. Furthermore, the data (up to 2018) reflects pre-reform and pre-pandemic patterns, and the illustrative value analysis is a sensitivity exercise rather than a direct policy prescription.

Technical Summary: Employment-type inequalities in healthy working life expectancy in China

More like this