The effect of urgent care centers on emergency department attendance and waiting times, with Peter Bristow and Luke Connelly (under review)
Lack of timely emergency department (ED) care can compromise patient outcomes, and ED timeliness has worsened across many health systems. One proposed response is diverting low-acuity patients from EDs via no-cost urgent care centres (UCCs). This study evaluates UCC introduction in Australia, where both ED and primary care are free at the point of use—unlike most existing co-payment-setting evidence, examining effects on ED activity, timeliness, and primary care utilisation. Using administrative data and a difference-in-differences design with dynamic treatment effects, we estimate a contemporaneous average treatment effect on the treated (ATET) for one additional UCC operating for one quarter, and examine heterogeneity by UCC–ED distance using cumulative exposure within radii of 1–15 kilometres. Across log, count, rate, and percentage specifications, with and without covariate adjustment, we find small reductions in semi-urgent and non-urgent presentations (generally 0–1% per quarter), somewhat larger but non-monotonic effects at shorter distances, and similarly small effects on overall ED demand. Timeliness measures show modest gains: up to a 1% increase in Category 3 presentations seen within recommended times and up to 0.4% more ED stays under four hours. Effects on general practice utilisation are small, with no evidence of an effect on mortality. Statistically significant estimates are not robust to plausible parallel-trends violations under Rambachan and Roth's (2023) sensitivity framework, across outcome definitions and specifications. Overall, UCC introduction has at most marginal effects on ED demand, performance, and primary care utilisation. Given estimated costs of approximately A$2.5 billion over seven years, UCCs alone are unlikely to materially reduce ED congestion.
The dynamic impact of parental separation on children's psychosocial development: Evidence from Australia, with Xiaoxue Dou, Bao Nguyen and Luke Connelly (under review)
Children’s emotional and behavioural development is a foundational determinant of their well-being and can have far-reaching implications for economic and social outcomes in adulthood, including social capital accumulation and labour market success. Whilst existing research has explored various determinants of children's emotional and behavioural development, the impact of parental separation remains ambiguous. This study contributes to the literature by exploiting the difference in timing of parental separation, employing a dynamic difference-in-differences framework, to estimate the short- and medium-term effects on children's emotional and behavioural development. Our findings suggest that parental separation has significant adverse effects on children's development, particularly in terms of hyperactivity, peer problems, and prosocial behaviour. The adverse effects peak immediately after separation and gradually diminish over time, though some subgroups, such as children from lower socioeconomic backgrounds or those whose parents had a good relationship before separation, experience more persistent negative effects. Our results underscore the importance of targeted interventions to support children during and after parental separation, particularly for vulnerable subgroups.
Motherhood and women’s health trajectories: evidence from Australia, with Xiaoxue Dou and Elcin Tuzel (draft available upon request)
We examine the impact of motherhood on women’s mental health and healthcare utilisation using longitudinal survey data from the Australian Longitudinal Study on Women’s Health linked to administrative records from 2000 to 2021. Exploiting variation in the timing of childbirth, we estimate dynamic treatment effects using the staggered difference-in-differences estimator. We find that motherhood leads to persistent deterioration in maternal mental health, particularly through increased depressive symptoms rather than anxiety disorders. The adverse effects intensify over time and are concentrated among women who have more than one child. The marginal effect of the second child seems particularly large. In contrast, first births generate relatively modest mental health effects. Administrative healthcare records reveal a striking divergence between mental health care needs and healthcare utilisation. Specialist mental health expenditure declines substantially following childbirth, particularly among mothers with multiple children, while antidepressant pharmaceutical expenditure increases persistently over time. These patterns suggest that worsening maternal mental health is accompanied by reduced use of specialist psychological services and increased reliance on medication-based treatment, likely reflecting time constraints, caregiving burdens, and barriers to accessing formal care. Additional analyses indicate larger mental health penalties among women with closely spaced births and among economically disadvantaged mothers.
Externalities of Large Public Sporting Events and Patient Outcomes in Nearby Hospitals, with Daniel Kuehnle and Rebecca Leber (work in progress)
Large sporting events attract substantial crowds and generate both positive and negative externalities. While prior research has focused on effects on spectators—such as disease transmission—little is known about how these events affect health outcomes for the broader population, particularly patients requiring urgent medical care. This project studies whether professional football matches in the German football league adversely affect patient outcomes admitted to nearby hospitals with sudden health shocks. Using high-quality administrative hospital data from Germany, we apply a difference-in-differences design comparing 30-day mortality for emergency admissions on matchdays versus non-matchdays. We focus on acute conditions such as heart attacks and strokes. The study provides new evidence on the health externalities of large sporting events and informs health service planning and public policy.
Wages and nurse turnover: Evidence from Australia, with Linn Htet Oo (draft available upon request)
This study provides new evidence on the relationship between wages and nurse turnover in Australia. Using data from the Household, Income and Labour Dynamics in Australia (HILDA) survey, we analyse 5,990 person-year observations of nurses from 2001 to 2023. Leveraging institutional features of Australia's nursing remuneration system, we estimate a linear probability model with two-way fixed effects and a carefully selected set of controls to mitigate potential bias from reverse causality and omitted variables. Our findings indicate that higher wages are associated with a lower likelihood of exiting the nursing profession: a 10\% increase in nursing wages corresponds to a 0.372 percentage point reduction in the probability of professional turnover. Nurse turnover appears notably higher in remote regions, aged-care settings, and the private sector, highlighting priority areas for targeted policy intervention and future research.
The impact of informal care provision on loneliness & social isolation, with Isaac Koomson and Raymond Kofinti (under review)
While informal care is regarded as an appealing alternative to formal care, it often comes at the expense of the caregiver. Existing studies have primarily focused on the financial implications of informal caregiving with limited attention to its hidden costs, including social aspects. This study contributes to the literature by examining the effects of informal caregiving on loneliness and social isolation using longitudinal data from a large nationally representative household survey in Australia. To study the relationship, we first employ a dynamic difference-in-differences (DID) method, exploiting the variation in the timing of caring across informal caregivers. Second, we complement our analysis with an instrumental variable (IV) method, using eligibility to government-funded Carer Payment as IV. Both strategies show that informal caregiving increases the probability of experiencing loneliness, particularly among female caregivers who provide care to their spouses. Our DID results do not show an effect on social isolation, whereas IV estimates suggest that informal caregiving increases social isolation among male compliers, but not females. Some of the effect of informal care on social health seems to be driven by changes to the relationship satisfaction between spouses when taking on a carer role.
Mass shootings and violent crime: The effect of subjective safety beliefs on health behavior, with Chris Rose and Brenda Gannon (draft available upon request)
Personal safety concerns consistently rank among the top issues in public opinion polls but are largely ignored in the economics literature on crime. This paper studies the impact of subjective beliefs around neighborhood safety on health behavior, namely physical activity, in the older adult population of the United States. We further provide new evidence of the reduced form impacts of mass shootings and crime on physical activity for this older group of people that have been overlooked in the literature so far, but who are disproportionately concerned about their safety. Using data from a large nationally representative survey of older adults between 2004 and 2020, we apply two complementary identification strategies. First, we exploit variation in timing and location of mass shootings and employ a staggered difference-in-differences estimator, showing that physical activity of respondents in the impacted counties declines in the years following the incident. Second, we implement an instrumental variables correlated random-coefficient estimator (IV-CRC), that allows for multi-valued treatments, catering for the categorical measurement of neighborhood safety beliefs. We use mass shootings and violent crime rates to construct instruments for safety beliefs and study their effect on physical activity. We find a significant deterrent effect of feeling unsafe on physical activity, which is stronger for females than males and more pronounced among people aged 75 and above.