How Much Do Income Shocks Drive Mental Health Declines After Losing a Spouse?

Fadlon, Fugleholm, and Nielsen (2025) ask whether income losses after a spouse’s death worsen survivors’ mental health. They use Danish administrative data from 1995–2018, including death records, prescription drug purchases, and household income information for the full population. They find that survivors sharply increase their use of mental-health medications after the death, with take-up doubling in the first month (about +10.5 percentage points) and remaining about 10 percent higher even four years later. They also show that income stability matters: a one standard-deviation smaller income loss reduces the first-year treatment effect by 7–11 percentage points, indicating that financial security protects mental health.

Why This Article Was Selected for The Policy Scientist

This article addresses an important and understudied policy question: the extent to which financial stability shapes mental health following spousal death. The topic is broadly relevant because population aging, widowhood, and household financial vulnerability intersect across advanced economies. It is timely given ongoing debates about social insurance design and mental health burdens. The authors have produced related work on household responses to health shocks, and this paper extends that line by linking income security to mental health. The use of comprehensive Danish administrative data strengthens internal validity, and the dynamic difference-in-differences design offers credible causal inference. The setting plausibly generalizes to other high-income jurisdictions with similar safety nets, though institutional features will vary.

Full Citation and Link to Article

Nielsen, T. H., Fugleholm, A. S., & Fadlon, I. (2025). Survivors’ mental health and the protective role of income stability. American Economic Review: Insights. https://www.aeaweb.org/articles?id=10.1257/aeri.20250006

Central Research Question

The article investigates the extent to which income stability mediates survivors’ mental health following the death of a spouse. The authors ask whether the mental health effects commonly observed after widowhood are driven solely by bereavement or whether financial losses play a substantively important role. This question situates mental health outcomes within a broader household economic framework and tests whether income shocks amplify psychological distress. More specifically, the researchers examine whether the reduction in household income triggered by a spouse’s death predicts increased psychotropic medication use, and whether income security—provided through public and private transfer systems—can offset these effects. The study therefore links mental health, mortality shocks, and household financial resilience within a single empirical framework.

Previous Literature

Prior research has consistently documented that spousal death is associated with adverse physical and mental health outcomes, higher health care utilization, and elevated mortality risk among survivors. Classical epidemiological and psychological studies have examined bereavement responses (e.g., depression, stress-related cardiovascular outcomes, suicide) and identified widowhood as a critical life event affecting both short- and long-term well-being. More recent work in economics has focused on health-related consumption, mortality spillovers, and insurance coverage, but rarely integrates mental health and household finances in a causal setting. The literature also includes evidence on health insurance expansions improving mental health and disability insurance payments reducing mortality. However, there is limited evidence on whether income security directly modifies mental health trajectories after sudden household shocks. The authors have previously contributed to this domain by studying family health behaviors and labor supply responses to health shocks, establishing their expertise in leveraging Danish registers for quasi-experimental identification. This article builds on that cumulative research by extending the analysis to psychotropic medication use and income losses after spousal death. While a subset of clinical studies has examined antidepressant prescriptions after bereavement, these studies have not linked effects to household economic mechanisms. The article therefore fills a gap between bereavement research in psychology and causal inference on financial shocks in economics, contributing a policy-relevant synthesis of two previously parallel research traditions.

Data

The authors utilize comprehensive Danish administrative data spanning 1995–2018. These data cover the universe of Danish residents and link individuals across registers containing mortality, hospitalization, pharmaceutical purchases, income, demographic characteristics, and family linkages. Deaths are identified from the national death register, while initial heart attacks or strokes are identified from hospitalization records, enabling the construction of clean onset events and measured survival horizons. Mental health outcomes are proxied using prescriptions for psycholeptic (N05) and psychoanaleptic (N06) medications, which capture a broad spectrum of anxiolytics and antidepressants. Income data derive from annual tax records and include labor earnings, disability insurance, old-age pensions, private pension payments, and insurance payouts. The authors define households as married or cohabiting couples aged 45–80 and construct income loss measures by comparing household income one year prior to the shock to one year after. The panel is balanced from four years before through four years after the shock to maintain comparability in dynamic analyses. These data provide unusually strong measurement precision relative to survey-based work and allow high-frequency outcome tracking at both monthly and annual levels. Because Denmark features universal health insurance and subsidized pharmaceuticals, mental health medication take-up reflects psychological needs more than affordability constraints, improving interpretation.

Methods

The researchers employ a dynamic difference-in-differences identification strategy comparing “treated” survivors—whose spouse experiences a fatal heart attack or stroke in a given year—to “control” survivors—whose spouse experiences the same type of fatal health event five years later. By choosing control couples that are destined to be affected by similar shocks, but at a later date, the authors construct a counterfactual trend representing underlying trajectories absent contemporaneous treatment. This design relies on the assumption that the precise year of a first heart attack or stroke within a limited horizon is quasi-random and not anticipated. Dynamic effects are estimated using household fixed effects, time fixed effects, and age controls. To examine heterogeneity in income losses, the authors instrument for the household’s income decline using the deceased spouse’s baseline share of household earnings, exploiting the fact that survivors lose more income when the deceased had higher pre-shock earnings. Instrumental variable specifications quantify how much of the treatment effect is explained by income loss rather than bereavement alone. The strategy yields dynamic event-study plots and two-stage least squares estimates that isolate financial channels. This design constitutes a credible causal inference approach and is preferable to multivariate regression without quasi-experimental structure. The absence of randomized assignment is unavoidable given the context, but the research design is robust and consistent with state-of-the-art observational causal strategies.

Findings/Size Effects

The authors find that spousal death generates immediate and substantial increases in psychotropic medication use. At the monthly level, take-up doubles in the first month following the death, rising by approximately 10.5 percentage points from a baseline of roughly 10 percent. At the annual level, medication take-up increases by 17.6 percentage points in the first year after widowhood, on a counterfactual baseline of approximately 25.6 percent. Effects decline gradually but persist for at least four years, with survivors still roughly 10 percent more likely to use psychotropic medication in year four. These patterns hold for both male and female survivors, with somewhat larger levels among women due to higher baseline rates.

Income stability significantly moderates these mental health effects. Survivors who lose an income-earning spouse exhibit larger immediate increases in medication use compared to those whose spouse had no earnings. Gender differences emerge due to differential labor force attachment: among households with deceased earners, females experience larger proportional increases in medication use than males. Instrumented regressions show that a one standard deviation smaller income loss (about 14 percentage points) reduces the first-year treatment effect by 7.8 percentage points for males and 11.4 percentage points for females. These magnitudes indicate that income losses explain a meaningful share of the observed mental health deterioration. Notably, the heterogeneity is concentrated in the first year, suggesting liquidity constraints rather than permanent income effects. Baseline income levels show only weak attenuation effects, and the presence of children does not substantively alter treatment effects. The results therefore imply that financial security—especially liquidity assistance—has protective effects on short-run mental health following spousal death.

Conclusion

The article demonstrates that spousal death generates large, immediate, and persistent mental health effects among surviving spouses, and that income stability mitigates a meaningful share of these declines. The findings link bereavement, household financial resources, and mental health in a causal framework, filling an important gap between epidemiological bereavement research and economic studies of income shocks. The study provides evidence that safety-net institutions that stabilize income can produce mental health benefits beyond their traditional consumption-smoothing role. Its contribution is strengthened by high-quality administrative data, precise measurement, and a credible quasi-experimental design. The external validity of the findings is highest for advanced economies with developed welfare states and comprehensive insurance systems, although the underlying mechanisms—income loss, liquidity constraints, and psychological stress—are plausibly general across institutional settings. The research advances the literature by establishing that financial channels matter for mental health responses to household mortality shocks and by quantifying effect sizes with policy relevance.