The transitory component of health care employment

Abstract We decompose U.S. health care employment into permanent and transitory components to study its cyclical properties over the 1990–2025 period. By isolating these short-run deviations from the long-run trend, we provide a robust framework to evaluate the ‘recession-proof’ hypothesis at the business-cycle frequency. We estimate a battery of linear and nonlinear specifications, explicitly accounting for the extreme volatility of the COVID-19 pandemic. Our results reveal significant model-based uncertainty: while we find strong evidence of asymmetric ‘bounce-back’ dynamics, alternative specifications yield divergent estimates regarding the timing, depth, and persistence of HCE fluctuations. To resolve this, we construct a model-averaged measure using an asymptotic approximation to the marginal likelihood, providing a consensus benchmark less sensitive to individual filter assumptions. We find that the HCE cycle is moderately procyclical, but this relationship is highly state-dependent and recession-specific. The sector exhibits ‘cyclical decoupling’ during periods of significant institutional or regulatory shifts, such as the 2008–09 financial crisis and the phasing in of the Affordable Care Act in 2014–15, suggesting that HCE possesses a conditional resilience driven by idiosyncratic factors and policy triggers rather than absolute immunity to general economic shocks.