Heterogeneity in the smoking response to health shocks by out-of-pocket spending risk

Abstract

An existing literature demonstrates that adverse changes to health can lead to improvements in health behaviors. Although the exact explanations for these empirical findings are debated, some posit that individuals learn about their true health risks through health shocks. Updated health risk information can then induce changes in health behaviors. While we follow a learning framework, we argue that past work has neglected the role of health insurance and medically related financial risk within this decision making context. Using longitudinal data from 11 European countries, we investigate the impact of a new cardiovascular (CV) health shock on smoking decisions among older adults and examine whether personal exposure to medical spending risk influences the smoking response. We then explore two potential mechanisms for this link: larger updates to health risk beliefs and higher medical expenditures to incentivize behavior change. We find that CV shocks impact the propensity to smoke, with relatively more impact among individuals with high financial risk exposure to medical spending. We also see larger increases in out-of-pocket expenditures following a shock for this group--consistent with the latter mechanism for behavior change.

Smoking Response to Health and Medical Spending Changes and the Role of Insurance

“…To explore heterogeneity in the smoking response, we focus on the degree of financial risk exposure to medical spending at the time of disease onset. Our preceding conceptual framework views self‐protection (i.e., smoking cessation) and market insurance as at least partial substitutes for a given smoker – consistent with other economics studies (Dave and Kaestner, ; De Preux, ; Ehrlich and Becker, ; Rashad and Markowitz, ; Richards and Helmchen, ; Richards and Marti, ). While health risk beliefs after a health shock could be similarly updated across insurance groups, the signal in the financial domain is likely stronger for individuals with less coverage.…”

“…The estimates in Table also indicated that NSR health problems could lead to higher medical expenditures but then failed to display any corresponding relationship with smoking behavior for either insurance group. Furthermore, Richards and Marti's () study finds a similar pattern of results among non‐US health care systems that typically establish OOP maximums for individuals – effectively eliminating the possibility of catastrophic losses and thereby large income effects. Our data also only encompass older adults, who are unlikely to be experiencing their first financial shock over the course of their smoking lifetimes.…”

“…The thought experiment supporting our primary analyses can be described as comparing two similar smokers with similar insurance status trajectories where one experiences the shock during a time of higher financial risk protection and the other experiences the shock during a time of lower risk protection. To improve identification relative to Richards and Marti (), we exploit two known features of the US health care system: a large uninsured population and the federally administered Medicare program.…”

“…In a recent paper, Richards and Marti () propose that a health shock may convey additional risk information in form of medical expenditures attributable to smoking. The authors then incorporate health insurance into the learning framework and examine the smoking response to health shocks by coverage levels.…”

“…Using European data, they find higher post‐shock cessation rates among smokers with lower financial risk protection and supporting evidence for medical costs as an underlying mechanism. We follow Richards and Marti () and argue that the onset of a severe health condition can have substantial financial consequences that are difficult for individuals to anticipate. However, experiencing a health shock provides credible and highly personalized risk information for both the health and financial risk domains – and possibly further incentive for behavior change.…”