The Effect of Health Shocks on Labor Force Participation for Older Adults: Have Patterns Changed Over Time?

Background: Health shocks, such as a new cancer diagnosis, acute myocardial infarction or stroke, have been shown to reduce labor force participation (LFP) in working adults, contributing to financial distress. Over the past decade, there have been economic swings, retirement age has increased, while newer treatments for some conditions may be less toxic or disabling. We update prior estimates of LFP responses, and assess whether patterns changed over time.

Methods: Adults from the 1998-2014 waves of the Health and Retirement Study (HRS) aged >50 years and working in 1998 or at HRS cohort entry in 2004 or 2010, were followed until a LFP decrease. Decreased LFP was defined as a switch from fulltime work, or a 10 hour or 25% decrease in hours worked. An alternative measure was limited to the more extreme response of leaving the LF. We created time varying indicators for health conditions (newly reported or worsened cancer, heart condition, stroke, lung disease) since the prior wave, condition history, demographics, household net worth, and insurance coverage. We estimated discrete Cox proportional hazard survival models for the full sample, with an interaction between observations pre/post 2007

Results: Among 13,771 working adults, 53.6% were <55 years of age, and 45.4% were women; health history at cohort entry included 5.3% cancer, 9.5% heart condition, 1.5% stroke, and 4.1% lung disease. Over the observation period, an additional 2.0%, 3.6%, 0.7%, and 2.1% of workers newly reported cancer, heart disease, stroke or lung disease, respectively. Health shocks were associated with increased risk of LFP decrease, with estimated effects ranging from a hazard ratio (HR)=1.69 (95% confidence interval (CI): 1.34-2.14; p<.0001) for stroke, to HR=1.17 (95%CI: 1.00-1.37) for lung disease. Prior history of heart disease, lung disease, diabetes, arthritis and psychiatric disorders were associated with LFP decreases, but prior history of cancer was not. Although decreased LFP was less likely post 2007, there was a differential effect only for new lung disease.

Conclusions: Health shocks in the form of new diagnoses or worsening of cancer and other serious health conditions were associated with LFP reductions, although a history of cancer was not associated with LFP changes. While population-level patterns indicate reduced rates of leaving the LF over time, there were no relative improvements in LF retention for older workers with health shocks. Reduction or loss of employment may contribute to loss of insurance and increased financial burdens. Specific strategies to support workers who have experienced health shocks and want to continue employment will be needed to further reduce LF effects in the current period.