📄 medicine

Impact of labor force participation on healthy life expectancy among older adults

Using CHARLS data and advanced statistical methods, this study demonstrates that labor force participation significantly extends both life expectancy and healthy life expectancy among older adults, with the magnitude of these benefits varying by baseline health, age, and urban-rural residence.

Original authors: Ruirui Xie, Shimin Su, Yiru Jiang, Tingting Xu, Huiling Dong, Bing-yi Wu, Zhenhua Feng

Published 2026-08-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ruirui Xie, Shimin Su, Yiru Jiang, Tingting Xu, Huiling Dong, Bing-yi Wu, Zhenhua Feng

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 around the world grow older, the question of how to live well in later years has moved from a personal concern to a global priority. For decades, the standard measure of success in aging was simply how long a person lived. Today, however, scientists and policymakers look at a different metric: healthy life expectancy. This concept asks not just how many years a person survives, but how many of those years are spent in good health, free from disability or chronic illness. It is a measure of quality, not just quantity. In societies where the number of older adults is rising rapidly, understanding what factors help people stay healthy and active for longer is essential. One of the most significant factors in daily life is work. For many, continuing to work after the traditional retirement age is a necessity; for others, it is a choice. But does staying in the workforce help an older person live a longer, healthier life, or does the physical and mental strain of work eventually wear them down?

A team of researchers in China set out to answer this question by looking at the lives of older adults between the ages of 65 and 85. They analyzed data from a massive, long-running survey that tracks the health and lives of thousands of Chinese citizens. The challenge they faced was that people who choose to keep working are often already healthier than those who stop. If a study simply compares workers to non-workers, it might mistakenly credit the work itself for the good health, when in reality, the good health allowed them to keep working in the first place. To get around this, the researchers used a statistical method to pair workers and non-workers who were otherwise very similar in age, education, and health history. Once they had these matched groups, they used a computer simulation to project how these people would age over time under different scenarios. This allowed them to estimate how many years of good health a person could expect if they continued working versus if they stopped.

The results revealed a story that changes as a person gets older. For adults in their mid-to-late sixties and early seventies, staying in the workforce was linked to a significantly longer life spent in good health. The simulation showed that a 65-year-old in good health who continued to work could expect to live for about 17 years, with roughly 11 and a half of those years spent in good health. In contrast, a similar person who stopped working was projected to live for about 13 years, with only about 7 and a half of those years in good health. This pattern held true even for those who started with poorer health; those who kept working still lived longer and spent a larger portion of their remaining years in good health compared to those who did not. The researchers suggest that the structure, social connection, and sense of purpose that work provides may help slow down the physical and mental decline that often comes with aging.

However, this benefit does not last forever. The study found that the protective effect of working begins to fade as people approach their eighties. After age 80, the advantage of continuing to work diminishes, and in some cases, the pattern reverses. For the oldest-old, the physical demands of work can become a burden rather than a benefit, potentially accelerating health decline. This suggests that the relationship between work and health is not a simple rule that applies to everyone at every stage of life. It is a balance that shifts as the body ages.

The study also highlighted a deep divide between city and country living. In both urban and rural areas, working older adults lived longer and healthier lives than those who did not work. However, the gap was particularly striking for rural residents. While city dwellers generally had access to better healthcare and resources, leading to longer healthy lives overall, the boost that working provided was even more dramatic for those in the countryside. A rural worker with good health at age 65 gained more relative years of healthy life from working than their urban counterpart did. This suggests that for older adults in rural areas, where resources are scarcer, the economic and social stability provided by work may be even more critical for maintaining health.

Ultimately, the research paints a picture of work as a powerful tool for healthy aging, but one that must be used with care. For younger seniors, staying active in the workforce appears to be a strong strategy for extending the years of life spent in good health. For the very old, however, the focus may need to shift away from economic productivity toward flexible social engagement and care. The findings do not suggest that everyone should work until they drop, but rather that the right kind of work, at the right time, can be a vital part of living a long and healthy life. As societies grapple with the realities of an aging population, these insights offer a clear path forward: support appropriate work opportunities for those who are ready, while ensuring that the oldest among us have the support they need to thrive without the strain of demanding labor.

Technical Summary: Impact of Labor Force Participation on Healthy Life Expectancy among Older Adults

Problem Statement
As China faces rapid population aging characterized by a large elderly base and low fertility, the metric for assessing population health is shifting from simple life expectancy (LE) to Healthy Life Expectancy (HLE). While labor force participation among older adults is a significant socioeconomic behavior, its long-term impact on HLE remains contested. Existing literature presents mixed findings, often describing a "double-edged sword" effect where work may either promote health through social engagement and economic security or accelerate decline due to physical strain and involuntary participation. Furthermore, empirical studies frequently suffer from sample selection bias (healthier individuals are more likely to work) and endogeneity issues, leading to potentially biased estimates of the causal relationship between work and health. There is a specific gap in understanding how these effects vary across different baseline health statuses, ages, and urban-rural divides over the long term.

Methodology
To address selection bias and estimate long-term health trajectories, the authors employed a two-stage analytical approach using data from the China Health and Retirement Longitudinal Study (CHARLS) waves 2018–2020. The final analytic sample consisted of 5,485 individuals aged 65–85.

Propensity Score Matching (PSM): To mitigate selection bias and endogeneity, the study first used PSM to construct comparable groups of labor force participants and non-participants. Propensity scores were estimated via logistic regression, and matching was performed using nearest-neighbor and radius (caliper=0.05) methods. This process balanced observed covariates (including age, sex, education, marital status, chronic diseases, and health behaviors) between the two groups, reducing the average treatment effect on the treated (ATT) bias.
Dynamic Microsimulation: Following PSM, a dynamic microsimulation model was applied to the matched sample to project individual health trajectories from age 65 to 85+. The model utilized multinomial logistic regression to estimate health-state transition probabilities (Healthy $\to$ Unhealthy, Healthy $\to$ Death, Unhealthy $\to$ Death) based on labor participation status and covariates.
HLE Calculation: The simulation accumulated time spent in the "healthy" state (defined as self-rated health of "very good" or "good") for each individual until death. The mean of these accumulated years across the simulated cohort provided the estimated HLE and LE under different scenarios (participation vs. non-participation, urban vs. rural, good vs. poor baseline health).

Key Contributions

Methodological Rigor: The study integrates PSM with dynamic microsimulation to address the dual challenges of selection bias and the need for long-term projection, offering a more robust causal inference than cross-sectional analyses.
Granular Stratification: Unlike studies focusing on aggregate effects, this research systematically analyzes heterogeneity across baseline health conditions (good vs. poor), age groups (specifically the 65–79 vs. 80+ divide), and residence (urban vs. rural).
Quantification of HLE Gains: The paper provides specific estimates of the absolute and relative gains in HLE associated with labor participation, moving beyond binary health outcomes to measure the quality of remaining life years.

Results

Overall Association: Labor force participation is positively associated with both LE and HLE. At age 65 with good baseline health, participants had an LE of 17.13 years and an HLE of 11.43 years (66.7% of remaining life), compared to 12.81 years LE and 7.62 years HLE (59.4%) for non-participants. Similar, though smaller, advantages were observed for those with poor baseline health.
Age-Dependent Effects: The protective effect of labor participation is age-dependent. Before age 80, participants showed lower probabilities of transitioning to unhealthy states or death compared to non-participants. However, after age 80, this trend reversed, with participants showing a higher probability of transitioning to an unhealthy state, suggesting that the benefits of work diminish or become detrimental for the "oldest-old."
Urban-Rural Heterogeneity: Urban older adults generally exhibited higher absolute LE and HLE than rural counterparts. However, the relative gain in HLE conferred by labor participation was more pronounced for rural older adults. For instance, among those with good baseline health, rural participants gained 2.16 more years of HLE compared to rural non-participants, a larger gap than seen in urban comparisons.
Baseline Health Interaction: The positive association between work and HLE persisted across both good and poor baseline health groups, though the absolute levels of HLE were naturally lower for those starting with poor health.

Significance and Claims
The authors claim that these findings highlight the non-linear and context-dependent nature of the labor-health relationship. They argue that labor participation is not uniformly beneficial but is highly beneficial for "younger-old" adults (65–79), likely due to structured routines, social engagement, and economic security. Conversely, for the "oldest-old" (80+), the physical demands of work may outweigh the benefits, leading to accelerated health decline.

The study suggests that policy should move beyond a binary view of "working vs. not working." Instead, it advocates for life-course-oriented approaches:

For younger-old adults, supporting appropriate, capacity-matched work opportunities can extend healthy longevity.
For the oldest-old, policy should prioritize flexible social engagement and care support over economically oriented employment to protect health.
Rural-specific interventions are necessary, as rural older adults with health limitations face cumulative disadvantages; integrating employment support with improved primary healthcare and occupational protection is crucial for reducing urban-rural health disparities.

The authors maintain a modest tone regarding causality, acknowledging that while PSM and microsimulation reduce bias, the observational nature of the data precludes definitive causal inference, and unobserved factors may still influence the results.

More like this