Fertility rates have declined globally, raising concerns about demographic aging and long-term economic sustainability. In response, many governments have introduced pronatalist policies to encourage childbearing. However, these efforts often overlook the psychological toll of raising children, especially in countries with limited institutional support for families. This study estimates the effect of having more than two children on maternal mental health using data from the IPUMS National Health Interview Survey (2000-2009). To address endogeneity in fertility decisions, we use the sex composition of the first two children as an instrument. We find that having more than two children increases maternal psychological distress by 0.69 standard deviations in the full sample and by 0.61 standard deviations among married mothers. Exploring heterogeneity, we find that effects are similar across income groups, while variation by maternal education is less consistent, with larger effects observed among married mothers with a high school education.
For Supplemental Nutrition Assistance Program (SNAP) households, the maximum benefit was temporarily increased by 15% from January through September 2021, with an April revision to Emergency Allotments (EA) raising benefits further for the lowest-net-income participants. We evaluate how the combined effects of these temporary benefit enhancements vary across households with different levels of susceptibility to food insufficiency, anxiety, and depression, as well as by household composition and geographic cost of living. Using data from the Household Pulse Survey, we first apply machine learning to predict households' susceptibility to each hardship outcome, then estimate treatment effects across the predicted susceptibility distribution using both traditional and doubly robust difference-in-differences estimators. The enhancements improved household food sufficiency by about 3%-4% on average. Gains tended to be larger among households with higher predicted susceptibility to food insufficiency and in lower-cost regions (e.g., 6%-7% in the Midwest), with smaller or null effects in higher-cost regions (e.g., the West). Across these dimensions, gains emerged largely after the April EA revision. We find no significant effects on anxiety over the January-September period in the full sample. In January-March, however, anxiety fell by about 3% across all households, driven by large reductions among those without children. Higher-cost regions also show anxiety reductions concentrated in this window. Depression shows no significant changes. We discuss implications for the design and targeting of SNAP benefit enhancements.
Two-thirds of children in low-income countries are affected by food insecurity. Household food insecurity is typically measured using caregiver reports, but children's reports may shed light on experiences or domains of food insecurity that caregivers overlook. Using a novel validated measure of child-reported food insecurity, we examine how both child- and caregiver-reported food insecurity predicted children's academic, cognitive, and social-emotional skills in a large sample of children (N = 3771, 5-17 years-of-age) and their caregivers in five northern regions of Ghana. Caregivers were nearly twice as likely as children to report food insecurity, with larger discrepancies across items reflecting moderate food insecurity manifestations. In value-added models that aim to control for unobserved heterogeneity that drives skills, caregiver-reported food insecurity predicted lower literacy and numeracy, especially among boys and older adolescents, whereas child-reported food insecurity did not predict any outcome. Potential mechanisms (child health, parental educational investments, children's aspirations, and caregiver psychological well-being) partially accounted for the observed associations. Findings suggest that caregiver reports better capture the constraints induced by food insecurity on children's learning in this context, which is important information for targeting interventions to support child development by tackling food insecurity.
International statistics reveal a persistent paradox: women receive depression diagnoses at rates two to three times higher than men, yet male suicide rates are two to three times higher than female rates. We investigate whether gender stereotype priming affects self-reported mental health using an experimental design embedded in a representative survey of the Italian population. We show that exposure to the treatment is associated to worse self-reported mental distress, with statistically comparable effects across genders. K-means clustering analysis reveals substantial heterogeneity: treatment effects are strongest among individuals who fail to conform to traditional gender expectations. These findings suggests that gender stereotype salience imposes psychological costs, particularly for those who fall short of traditional role expectations, offering insights into persistent gender disparities in mental health reporting and outcomes.
As populations age, many countries have increased the statutory retirement age to sustain pension systems and maintain labor supply. This article examines whether delayed retirement increases workplace accident risks, a potentially important but underexplored cost of pension reforms. We exploit the 2012 Italian pension reform - which sharply tightened eligibility requirements and abruptly raised the retirement age - to estimate its causal effects using administrative data on private-sector employees. Leveraging firm-level variation in exposure to the reform, we identify the impact of delayed retirement on both older workers compelled to remain employed and their coworkers. The reform led to a modest but statistically significant increase in sick leave due to occupational injuries equivalent to 2.6% of the pre-reform mean. This effect mainly reflects the direct impact on retained older workers, with no evidence of spillovers to coworkers. While the aggregate effect is limited, the adverse consequences are highly unevenly distributed, being concentrated in small firms (fewer than 15 employees) and in sectors with high physical or cognitive demands, such as manufacturing and professional or technical services. Overall, our findings highlight occupational safety as a potentially hidden and distributional cost of pension reforms aimed at extending working lives.
Large-scale ecological restoration is increasingly promoted as a strategy for environmental sustainability, but its implications for population health remain underexamined. We study whether China’s national ecological restoration pilot improved residents’ self-rated health, and through which pathways these effects emerged. Using six waves of the China Family Panel Studies and the staggered rollout of the pilot across prefecture-level cities, we estimate two-way fixed-effects difference-in-differences models linking policy exposure to individual health outcomes. Ecological restoration was associated with a 0.1697-point increase in self-rated health on a five-point scale. Event-study estimates showed no evidence of differential pre-policy trends, and results were robust to placebo tests, Goodman-Bacon decomposition, propensity score matching, double machine learning, instrumental-variable analyses, and adjustment for overlapping environmental policies. Mediation analyses suggest that the health gains operated through environmental and biophysical, economic and material, and psychosocial pathways, including lower PM₂.₅ concentrations, greater greenness, reduced chronic disease, longer sleep duration, more frequent exercise, improved visibility, lower income inequality, higher perceived socioeconomic status, greater life satisfaction, and higher household income. Policy exposure was also associated with lower household medical and preventive-health expenditures. These findings suggest that place-based ecological restoration can function as a population health intervention by reshaping environmental exposures, material resources, and psychosocial conditions.
After the global financial crisis of 2007-2008 and the subsequent Great Recession, many governments in European Union countries implemented fiscal consolidation programmes aimed at restoring public finances, often through expenditure cuts. We examine the impact of these programmes on a broad set of health indicators capturing overall population health as well as vulnerabilities at early and premature stages of the life cycle. Exploiting the staggered adoption of fiscal consolidation programmes across countries, we estimate a Difference-in-Differences model with multiple time periods and variation in treatment timing. Using data for the 27 European Union countries over the period 1995-2015, we find that the fiscal consolidation programmes are associated with higher mortality and premature mortality, increased infant mortality, and lower life expectancy.
This paper examines the decongestion effect hypothesis by exploring whether higher private healthcare financing is associated with better access to public services, as would be expected if private financing were linked to lower pressure on the public healthcare system. Using data from Italian regions over the period 2017-2022, we consider self-reported unmet needs for specialist visits as an indicator of barriers to access to public services, distinguishing between financial and waiting-time constraints. We estimate a bivariate probit model to analyze the association between regional private healthcare expenditure and unmet specialist visits, controlling for individual socioeconomic characteristics and regional health system features. The findings indicate a positive association, particularly among less affluent individuals. In contrast, higher regional public spending is associated with lower levels of forgone care. Overall, the results do not provide empirical support for the view that greater private financing is linked to better access to public healthcare services. They also raise questions about policy interventions, such as tax incentives for voluntary health insurance, that promote private healthcare financing as a tool to improve the capacity of the public healthcare system to meet patients' needs.
Despite growing research interest in school entry policies, their effects on newborn well-being remain underexplored. This study provides new causal evidence using restricted birth data from three states, based on the 2003 revision of birth certificate instruments. Our normalized-and-pooled regression discontinuity analysis shows that young mothers born just after the school entry cutoff have significantly lower levels of education than those born just before. These women also tend to experience poorer birth outcomes, including a modest reduction in birth weight and higher risks of adverse pregnancy outcomes, though the strength and precision of these estimated effects vary across states. Maternal health behaviors, insurance coverage, and paternal age emerge as important channels for these intergenerational health effects.
This study examines the impact of South Korea’s Basic Pension expansion in 2014 on social isolation, focusing on social connectedness, loneliness, and depressive symptoms among older adults. Using data from the Korean Longitudinal Study of Ageing (KLoSA), this study employs a Difference-in-Differences (DiD) approach combined with an Instrumental Variable (IV) strategy, which leverages the pension eligibility threshold at age 65 and the 2014 policy expansion. The results show that the pension expansion significantly reduced social isolation. After 2014, Basic Pension recipients reported more frequent social interactions, lower loneliness, and fewer depressive symptoms. The IV estimates confirm a causal relationship, suggesting that greater financial security facilitates social participation and alleviates loneliness and depression.
This paper examines the long-term effects of early-life exposure to county health departments (CHDs), a public health initiative aimed at providing affordable healthcare access to disadvantaged and underserved areas during the early decades of the 20th century, on later-life longevity. We employ Social Security Administration death records linked with the 1940 census and implement a difference-in-differences estimation strategy to compare the longevity of individuals exposed at different ages to the county-specific year of CHD establishment. We find that exposure to a CHD opening in the county of birth during fetal development and early years of life is associated with a 1-4-month increase in longevity. We observe considerable heterogeneity in the results, such that the effects are more concentrated among non-white individuals and individuals with lower-educated parents. Contemporaneous fertility and infant mortality data suggest reductions in infant mortality rates and increases in birth rates following CHD openings. Additionally, we find significant improvements in educational attainment, socioeconomic status, and income in early adulthood. We argue that enhancements in the fetal environment, improved fetal survival, and better early life health capital resulting from CHD exposure contribute to improvements in human capital during adulthood, with potential implications for later-life longevity.
We study the extent to which exposure to COVID-19 relates to healthcare professionals’ prosociality. Drawing on empirical evidence from an incentivized experiment and a companion survey of healthcare professionals (HCPs) at a large Italian hospital (N = 194), we find that different forms of exposure to COVID-19 predict their altruistic motivation in heterogeneous ways. HCPs who worked in COVID-19 wards or had a close relative or friend severely affected by the virus are more than 5 percentage points more likely to prioritize patient welfare over personal gain, with the association for professional exposure being mostly driven by female HCPs, especially nurses. Conversely, personally contracting COVID-19 is associated with a 6 percentage point decline in prosociality. Our results highlight that different experiences of need shape prosocial behavior in hospital settings.
Does public health insurance mitigate the functional consequences of adverse health shocks? I investigate this question by exploiting variation induced by Tennessee's 2005 Medicaid disenrollment, which abruptly terminated coverage for approximately 170,000 childless adults. Using Behavioral Risk Factor Surveillance System (BRFSS) data from 1997 to 2010 and a difference-in-differences design, I estimate Intent-to-Treat effects. I first document that the reform significantly increased the uninsured rate and increased the share of Tennessean childless adults reporting "poor" health status. I also find suggestive evidence of an increase in severe obesity prevalence. I then document a puzzle: while reported days of functional incapacitation increased by 27%, the underlying frequency of physical or mental unhealthy days remained statistically unchanged. I reconcile this divergence by documenting suggestive evidence that the "incapacitation ratio" - the probability that an unhealthy day results in incapacitation - increased by 17%. Further evidence suggests that reduced access to care is a plausible mechanism: following the TennCare disenrollment, the share of childless adults unable to see a doctor due to cost increased by 2.5 percentage points (23%), the evolution of which tracks the evolution of incapacitation ratio well. Taken together, the findings suggest that public health insurance may preserve health resilience, likely by mitigating financial barriers to care.
We examine the effects of prenatal exposure to fine particulate matter (PM2.5) on birth outcomes in a low-pollution setting. We combine administrative birth records in Quebec between 2008 and 2015 with air quality monitoring data and linked tax records in h neighborhood-by-quarter fixed effects models. While we find little evidence that prenatal PM2.5 exposure substantially affects birth outcomes on average at the population level, we unmask important heterogeneity. Constructing a cumulative socioeconomic vulnerability index, we find that in most vulnerable families a 1μg/m3 increase in average PM2.5 exposure is associated with a reduction of approximately 6.6 grams in birth weight. Focusing on a sample of births admitted to one of the province’s neonatal care unit, we also find important effects of in utero PM2.5 exposure on biologically vulnerable newborns, both in terms of prematurity and of birth weight. We also find that days exceeding the Canadian 24-hour PM2.5 standard have disproportionately large effects among vulnerable families, such that peaks in exposure matter beyond average exposure itself.
This paper studies the effect of mental health on employment and labour-force participation in China using nationally-representative China Family Panel Studies (CFPS) data. Utilising CFPS waves 2012, 2016, and 2018, we construct consistent measures of labour supply and mental health using the Center for Epidemiological Studies Depression (CES-D) scale. The panel structure enables us to control for unobservable time-invariant individual heterogeneity, in contrast to most prior studies from developing countries. To address concerns about reverse causality and omitted time-varying factors, we complement fixed-effects estimates with (i) Oster (2019) coefficient-stability bounds and (ii) robustness checks using respondents' reported reasons for non-employment. Our estimates reveal that experiencing severe depressive symptoms reduces the likelihood of employment and labour-force participation by, on average, 1.8-1.9 and 1.3-1.4 percentage points, respectively. These estimates are similar in magnitude to those from advanced economies, despite China substantially differing in terms of its development, labour market structure, and mental health provisions. Within China, the estimated effects are concentrated almost exclusively among older male workers, with limited evidence of comparable impacts for the broader prime-age workforce.
Waiting time for elective treatments is a major health policy issue in many OECD countries that was exacerbated by COVID-19. We test whether waiting times from referral to surgery reduces post-operative health and resource use for patients in need of a hip or knee replacement in England. We employ a large administrative dataset for 2015/16-2022/23. We use an instrumental-variable regression approach based on congestion while controlling for a rich set of patient characteristics and hospital fixed effects. We find that, for public hospitals in the preCOVID period (April 2015-January 2020), referral-to-treatment waiting times do not have a statistically significant effect on post-operative health for either hip or knee replacement (as measured by Oxford Hip or Knee Score) or the probability of being readmitted within 28 days following discharge. However, for public hospitals, one additional month of waiting instead increases length of stay by 0.075 days (2.06%) for hip replacement but has no effect for knee replacement. For private hospitals, we do not find an effect between referral-to-treatment waiting times and any of the outcomes. In heterogeneity analysis, we generally do not find differential effects of waiting on different groups (by gender, ethnicity, income and education deprivation, pre-operative health). In the COVID-19 period (April 2020-March 2023), we generally do not find an effect of waiting times on any of the outcomes.
A gap in mortality rates between rural and urban areas emerged in the late 1990s as rural residents began dying at a higher rate than their urban counterparts. This mortality gap has been widening ever since. The growing mortality gap is pronounced among prime working-age adults (ages 25-54), which has important implications for rural health, productivity, and economic development. Despite the heightened mortality rates in rural areas in recent decades, there is still a limited understanding of the health-related mechanisms fueling this worsening rural-urban mortality disparity. Using nationally representative data from the 1999-2000-2017-March 2020 cycles of the National Health and Nutrition Examination Survey, we combine individual-level data, including biomarkers, and place-based characteristics to create a comprehensive dataset spanning 20 years. In a descriptive analysis, we use linear regression models to estimate the magnitude of rural-urban gaps in health behaviors and outcomes-or mortality risk indicators-for the overall adult and prime working-age populations. We document a suite of rural health disadvantages, which may be informative for understanding the growing rural longevity gap. These disadvantages are often attenuated and no longer statistically significant after accounting for county-level characteristics. A complementary decomposition analysis indicates that county characteristics account for a larger share of the variation in health measures than rural-urban status. Our results can inform decision-makers aiming to improve rural health and economic outcomes and may spur further research using biomarker data alongside place-based characteristics.
Stunting remains a major development challenge in many countries, including Ecuador. In this paper, I examine whether agrarian reform policies implemented during the 1960s and 1970s help explain long-run patterns of child stunting. The Ecuadorian reform relied on two distinct land allocation strategies: Public land transfers (PLT), designed to promote frontier settlement, and expropriation, aimed at redistributing land from large estates. Combining household survey data with historical maps and administrative records, I estimate the long-run relationship between these policies and child stunting by exploiting historical variation in land allocation across parishes. I find that areas exposed to PLT exhibit significantly lower rates of child stunting, while expropriation shows no robust association. Across specifications, PLT exposure is associated with stunting rates that are 3-17 percent lower relative to the mean. Cohort evidence further indicates that the relationship is stronger for mothers plausibly exposed to PLT during early childhood. Consistent with this pattern, mothers from more exposed PLT cohorts attained higher levels of education, suggesting that maternal human capital may be one channel linking frontier settlement policies to lower child stunting.
Who falls into suicidal ideation when income drops? Drawing on the Korea Welfare Panel Study covering 2011 to 2023 and decomposing income into permanent and transitory components, I find the answer is highly concentrated. Transitory income fluctuations leave ideation almost unchanged, while permanent declines are associated with strong and durable increases in ideation, extending the permanent income logic to mental health. Households suffering persistent income declines and renters or the initially poor bear nearly all of the cost: renters carry a gradient roughly five times that of homeowners, and the initially poor several times that of the better-off. Material hardship and depressive symptoms together absorb the bulk of the income-ideation link. Administrative regional records on suicide deaths reproduce the negative income gradient and reveal a pooled male mortality response about six times the female one, even though ideation itself is more prevalent in women, a layered cognition-behaviour pattern with sex-symmetric thoughts but male-predominant lethality. The income-to-ideation gradient is markedly steeper among heavier problem-drinkers: an alcohol-disinhibition effect that operates across both sexes at the cognitive level, while the step from ideation to lethal behaviour falls predominantly on men. Together the results identify housing tenure and initial deprivation as the pressure points along which the income-mental-health gradient bites, and where social protection has the greatest scope to act.
