My research
For a link to my Research Statement click here.
Job Market Paper
When divorce is a credible threat
Abstract: Considering the socio-economic advancement of females over the past 50 years, this article reimagines the collective model of household bargaining to consider higher-wage partners relative to lower-wage partners, superseding the traditional husband/wife dichotomy. This article relies on a predicted higher-wage/lower-wage dichotomy, as well as providing results for the actual higher-wage/lower-wage dichotomy, to address concerns that wages are endogenous to the household bargaining process. Furthermore, over the past 50 years jurisdictions around the world have been adopting laws that empower females in the context of relationships, such as unilateral divorce. This article utilises unique survey data recording attitudes towards kinship such as `Marriage is an outdated institution'. Responses to these questions are used to provide new empirical evidence on the intersection of preferences, social norms and the law. Attitudes that are permissive towards divorce have the same effect on household bargaining dynamics as laws that are permissive towards divorce. The data is interpreted using a collective model of household bargaining with no inter-temporal commitment.
For a link to this paper in DropBox click here.
Working Papers
Societal implications of flexible work arrangements and environmental shocks: an allostatic load perspective (with Russell A. Matthews, Michael K. Price, Laura Razzolini and Madeline Wertschulte)
Abstract: Increasingly frequent and severe 'complex emergencies' have prompted calls for policy-orientated mechanisms for enhancing resiliency as a proactive way to prepare for disasters and individual vulnerabilities. Within the public policy space, flexible work arrangements have emerged as a prime way to proactively intervene, at a structural level, to promote resilience. The current program of research applies allostatic load theory to frame an events-based examination of flexible work arrangements as means to promote resistance and resiliency in response to two archetypical environmental shocks (i.e., commute strain, perceived infection risk). Utilizing data from three distinct longitudinal surveys, we conducted four studies employing a two-stage least squares analysis approach. The first-stage estimates the impact of environmental shocks and moderators (e.g., infection risks and flexible work arrangements) on allostatic overload (e.g., stress or SF-36). The second-stage then examines the impact of the toxic environmental shock and flexible work arrangements on tertiary outcomes of interest (e.g., job satisfaction) via their impact on endogenously determined factors (e.g., allostatic overload). In support of our structural intervention argument, flexible work arrangements moderated the impacts of shock on allostatic load. And as predicted, allostatic load negatively impacts tertiary outcomes. Furthermore, we show that flexible work arrangements have an ‘immunization’ effect even in the face of these shocks. The evidence surrounding the positive impact of flexible work arrangements on resilience in response to toxic environmental shocks has clear implications for building resilience through government policies with far-reaching social benefit. Practical contribution and broader impact findings are discussed.
A draft of this paper can be found here.
Locus of Control and Pro-Social Behavior (with Mark A. Andor, James Cox, Andreas Gerster, Michael Price, Stephan Sommer and Lukas Tomberg)
Abstract: We investigate how locus of control beliefs – the extent to which individuals attribute control over events in their life to themselves as opposed to outside factors – affect prosocial behavior and the private provision of public goods. We begin by developing a conceptual framework showing how locus of control beliefs serve as a weight placed on the returns from one’s own contributions (impure altruism) and others contributions (pure altruism). Using multiple data sets from Germany and the U.S., we show that individuals who relate consequences to their own behavior are more likely to contribute to climate change mitigation, to donate money and in-kind gifts to charitable causes, to share money with others, to cast a vote in parliamentary elections, and to donate blood. Our results provide comprehensive evidence that locus of control beliefs affect prosocial behavior.
An earlier version of this article was released by my co-authors as an NBER Working paper, located here. The purpose of me joining his project was to utilise the Household, Income and Labour Dynamics in Australia (HILDA) Survey to demonstrate that the effect of locus of control on pro-social behaviour could not be explained by correlation between locus of control and intelligence (as proxied by three tests) or the Big 5 personality traits. I also provided further external validity to this paper’s results.
Hours mismatch, parenthood and work-life balance (with Michael K. Price)
Abstract: We find that hours mismatch, either under-employment or over-employment, has differential impacts on parents’ perception of work-life balance, dependent on the age of their children. Parenting can be alternately cost-intensive or time-intensive as children age, which affects the extent of disutility a parent feels from hours mismatch. We utilise quantile regression to reveal that parenthood has a more significant impact on both under-employment and over-employment than previous research. “Quiet quitting”, especially since the COVID-19 pandemic, is becoming more common. Workers are now willing to trade income for a better work-life balance. Furthermore, worker well-being affects firm productivity. This research provides a deeper understanding of the effect of hours mismatch on individuals’ well-being over the life cycle.
This article extends a chapter of my thesis, located here. Prof. Price saw promise in this original paper but did not see it becoming a leading publication. We are now co-authoring the paper and incorporating considerations of work-life balance into the analysis.
Socio-biological determinants of gender, as revealed through time use
Abstract: Considering progressive gender attitudes prioritise the needs of females, and traditional gender attitudes prioritise the needs of males, it is puzzling that I find the sexes gender attitudes hardly differ. I utilise an index from biological anthropology, Gendered Fitness Interest (GFI), that calculates the sex-bias of how individuals' genes will be passed on. If a female (male) has many female descendants of reproductive age they take actions reflective with internalising a more progressive gender attitude and in turn devote more (less) time to labour market work. If a female (male) has many male descendants of reproductive age they take actions reflective with internalising a more traditional gender attitude and perform more (less) non-paid work. Therefore, an individual may act in accordance with attitudes that do not favour their own needs, but favour the needs of their descendants. I conceptualise gender as dynamic and lifelong; gender is an action, revealed in the sexual division of labour within the home. This research helps explain why, in social science research, some individuals appear to have preferences that do not prioritise their interests.
This article is a chapter from my thesis, located here.
The motherhood penalty, household means testing and student loan debt
Working Abstract: This article examines how a hybrid system of individual income taxation and household-based means-testing, as well as an income-contingent loan system for higher education, shapes gendered disparities in student loan debt. By focusing on whether mothers, more than fathers or childless females, are likely to hold outstanding student loan debt, the analysis explores a previously underexamined dimension of the motherhood penalty. This ‘motherhood penalty’ in student debt is exacerbated by a tax and transfer system which discourages dual-earner households by withdrawing benefits based on combined household income. Mothers, more likely to reduce labour force participation after childbirth than fathers, often remain below the repayment threshold for student loans, leading to debt accumulation over time. The findings have important implications for student loan policy, family benefit design, and the broader economic consequence of motherhood on female’s financial trajectories.
For a two-page Pre-Analysis Plan of this research, please click here.
Projects in the Pipeline
A National Disability Insurance Scheme and Household Inequality
Working Abstract: The National Disability Insurance Scheme (NDIS) was launched in Australia in 2013 to provide more extensive and coordinated supports to individuals who experience significant disability. The NDIS is the second most expensive government program in Australia, after the aged pension. Given this significant fiscal cost, it is important to evaluate the schemes effectiveness. Whilst the effectiveness of the NDIS could be evaluated using many different metrics, I specifically propose to analyse how the NDIS has affected household inequality. I propose to utilise the Person-Level Integrated Data Asset. This is a rich administrative data source that contains information on NDIS participants since the program’s inception, as well as demographic, income, welfare and health data for the years 2006-2023. Furthermore, PLIDA enables persons to be linked to dwellings, from which one can use the OECD-modified equivalence scale to measure household equivalised income. In the instance of disabled children or severely disabled adults, these individuals will not be able to work even with significant supports. However, the supports provided by the NDIS should facilitate the employment of parents, partners and other carers of NDIS participants. I intend to exploit two natural experiment methodologies. First, I propose a staggered differences-in-differences design calculating the gap in household equivalised income between NDIS participants and non-participants over the five year period pre-2013 with the 10 year period post-2013. Second, I propose a regression discontinuity design. Persons become eligible for the NDIS at the age of 9 and ineligible at the age of 65, however schooling usually starts at age 5 in Australia and the aged pension eligibility is for those aged over 67. I intend to look at how household equivalised income changes for NDIS participants at both age cut-offs. Providing evidence on how the NDIS has affected household inequality could provide information as to whether other countries should adopt a similar scheme, as well as inform public discourse in Australia where many now say the scheme is “too generous” and should be “rolled back”. With time, I intend to investigate a broader range of outcomes of the impact of the NDIS, rather than simply analysing household inequality. In line with current trends in the economic literature, I intend to look at mediating effects and moderating effects of this impact evaluation, to ascertain under which circumstances the program is scalable. This could help inform the better target and design of programs such as the NDIS to enhance their impacts. For example, the NDIS could increase household equivalised income through helping prevent relationship breakdown (a mediating factor). Perhaps the NDIS may only be effective in largely populated areas (a moderating factor). Another consideration may be that the NDIS is primarily accessed by wealthier and better educated individuals, who have the social capital necessary to navigate government services with complicated eligibility requirements.
This article intends to utilise my significant professional experience at the Australian Bureau of Statistics with the Person-Level Integrated Data Asset and other administrative data. For a two-page Pre-Analysis plan of this research, please see click here.
Nature, Nurture and Intergenerational Earnings Elasticities
Working Abstract: The correlation in earnings between fathers' and sons' earnings, which has fluctuated between strong, moderate and weak over the past century, is an important metric of intergenerational economic mobility. I utilise unique survey data, the Household, Income and Labour Dynamics in Australia Survey 2001-2023 (HILDA), to investigate whether a strong correlation between fathers' and sons' occupational status and earnings is due to nature or nurture. To do so, I utilise two different identification strategies. First, for all respondents, HILDA provides an occupational status scale of an individual as well as both their parents. The data also records if an individual’s parents were divorced at the age of 14, and which parent they primarily lived with. I find that there is greater correlation between fathers’ and sons’ occupational status if an son lived with their father at the age of 14. Furthermore, cohort effects show this occupational status elasticity has been growing stronger recently, indicating that intergenerational economic mobility is declining. My second identification strategy is to look at children of divorced parents at the start of the HILDA survey. HILDA records the number of years a child has had a stepfather, and the number of days per month they usually reside in the same house as the stepfather. I find that the earnings of step-children at 30 are more highly correlated with the earnings of step-fathers the greater the exposure the child had with the step father. This is evidence that nurture, as well as nature, is an important explanatory feature of intergenerational earnings elasticities. This research is intended to help inform intervention strategies to increases economic mobility for poorer- background children.
Working paper coming soon.
Social Isolation and Mortality: A population-level study
Working Abstract: In 2023 the US Surgeon-General called attention to an “epidemic of loneliness and isolation”. He released a report outlining how the physical health consequences of poor or insufficient connection include a 29% increased risk of heart disease, a 32% increased risk of stroke, and a 50% increased risk of developing dementia for older adults. Additionally, lacking social connection increases risk of premature death by more than 60%. I propose using a unique combination of Census and administrative data to undertake a population-level study of how time spent in social isolation can lead to early mortality. I intend to use the Person-Level Integrated Data Asset (PLIDA) which is an Australian administrative data source. PLIDA records date of death and cause of death. In contrast, it is frequently difficult to distinguish death from attrition in survey data. Crucially, PLIDA enables persons to be linked to dwellings, so I am able to ascertain if a person lives alone, and if so, for how long. Second, PLIDA is geo-spatially enabled. PLIDA divides Australia into hierarchies of geographic areas where the ‘population centroid’, or densest mass of population for each hierarchy and geographic area, is given a longitude and latitude. Therefore, I can calculate the distance between a person’s home (which is also given a latitude and longitude) and a population centroid. There is significant variation in individuals’ distances to population centroids in Australia, as Australia is a large yet sparsely populated country. I intend to utilise a Cox Proportional Hazard model, stratified by age, where the key independent variables are the length of time a person has lived alone, their distance to a population centroid, and the interactions of these variables. Crucially, PLIDA provides extensive health record data from which I can proxy ease of access to, and use of, health services. This study is intended to create a factual basis for public policy initiatives to decrease social isolation. This article intends to utilise my significant professional experience at the Australian Bureau of Statistics with the Person-Level Integrated Data Asset and other administrative data.