Labor supply effects of ill health: a weighted instrumental variable approach to misclassification of health measures

This paper examines the impact of ill health on labor supply, addressing a key challenge in
economic research: the misreporting of self-assessed health status. To overcome limitations in
prior empirical methods for handling misclassified health, we propose a new approach that
identifies the true effect of ill health using observations less likely to be misreported. Rather than
assuming fixed misclassification rates, we develop a semi-parametrically estimated health index
that extracts more accurate health information from available data. By leveraging this health index
to inform misreporting probabilities, we employ a weighted IV estimator and optimize the
weighting scheme to balance the tradeoff between squared bias and variance. We demonstrate the
superior performance of our model via both simulations and real-world data. Using data from the
2012 Health and Retirement Study (HRS), our findings indeed suggest that conventional methods,
including OLS and standard IV techniques, significantly underestimate the negative impact of ill
health on labor supply. Our approach reveals a much larger reduction in labor market participation
due to ill health, highlighting the economic vulnerability of individuals with health limitations.
These findings have important policy implicationsfor designing social safety nets and employment
policies to better support workers facing health challenges.