| YEAR | GRADUATE | JOB PLACEMENT | Thesis | Abstract |
|---|---|---|---|---|
| 2024 | Thomas Bulman | TBA | What is supermarkets’ influence on food waste within the food industry? | Food waste is a significant environmental, social, and economic challenge that spans the entire food supply chain. Although academic literature predominantly focuses on measuring and mitigating food waste at the consumer level, this emphasis leaves a gap in understanding the causes of food waste within the food industry and throughout the supply chain. This paper aims to address that gap by investigating whether supermarket power influences food waste in the food industry and quantifying the magnitude of this effect. Using food waste data from U.S. states from 2011 to 2021, this study is among the first to analyse the relationship between market power and food waste empirically. Applying a Generalised Method of Moments model shows that the impact of market power on food waste depends on price, represented by 10.3−9.4× (Price Change), indicating that as market power increases, total food waste also rises. This paper contributes to the literature by demonstrating that market power positively influences food waste and emphasises the need for further research to clarify the channels through which market power leads to waste, informing the development and implementation of effective policies to reduce food waste. |
| 2024 | Dominic Furneaux-Bate | Senior Research Officer - NSW Bureau of Crime Statistics and Research | Game Day: The Impact of National Rugby League Games on Domestic Assault in New South Wales | I investigate the causal relationship between National Rugby League (NRL) games and domestic assaults (DA) using panel data from the Bureau of Crime Statistics and Research (BOCSAR) from 2017 to 2023. I analyse regular season Newcastle Knights games and the more salient State of Origin series. I leverage the geographical isolation of the Greater Newcastle area from Sydney to introduce variation in regular season games. I employ a two-way fixed effects (TWFE) event study and an interaction-weighted estimator (IWE) to account for heterogeneous treatment timing and effects, capturing the dynamics of NRL games over a 24-hour window. I find that Newcastle Knights games increase DA by 8.4%, while State of Origin games lead to a statistically significant 29% rise. I explore the mechanisms driving these effects, finding that alcohol is the primary factor; I do not find evidence that emotions significantly affect domestic assault dynamics. My findings provide a foundation for future research and targeted policies to address the impact of sporting events and alcohol consumption on domestic assault in Australia. |
| 2024 | Billy Norris | Graduate Economist - Oxford Economics Australia | How Partisan Are Australian Newspapers During Federal Elections? | The potential impact of the media on human judgement in Australia, paired with a competitive two-party system, warrants an investigation into levels of partisanship within the Australian press. When an Australian election occurs every three years, the outcomes decide who is in power, thus significantly influencing the macro-economy and government spending. Hence, the media has room and incentive for potentially swaying Australian voters and shaping the nation’s future. Using data from The National Library of Australia I estimate partisanship across 22 major newspapers over five election periods: 2010, 2013, 2016, 2019, and 2022. This paper reports findings derived from collecting 2,277 newspaper headlines and articles. After employing both human and AI coders to score articles’ slant, I use Fixed Effects Regressions to calculate average slant ratings for all data points and groups within the paper. I also map slant over time, both individually and grouped by parent company. My findings indicate that Australian newspapers were clearly split between Labor and Coalition, with the majority of Rupert Murdoch’s News Corporation papers being slanted right towards the Liberal-National Coalition (LNP), while Fairfax media was slanted left towards the Australian Labor Party (ALP). However, the partisanship positioning for all newspapers has not been consistent, with major deviations in the 2013 and 2019 elections. These results imply that, on average, partisanship is not stable over time, potentially impacting voters’ ability to make balanced decisions during the selected election periods. |
| 2024 | Samuel Potter | TBA | How do participants in financial markets value information? | In seeking to research how participants in financial markets value information, we modify an existing artificial market model to provide us with the tools to do so. Taking the self-referential Sante Fe artificial market model, and modifying how agents form expectations, we take a crucial step towards answering questions regarding the value of information in markets, setting up potential future research of the topic. We also investigate potential improvements in the running of experiments with this model, seeking to shorten the time spent running simulations while still producing useful results. We analyse the results of these modifications and make suggestions for further modifications for potential future research. We conclude that the modification was successful and can provide the basis for future study of the value of information. |
| 2023 | Aaron Bannerman | Equities Customer Consultant - AUSIEX | Perceived housing wealth effect on young adult's co-residence | This thesis examines how perceived housing wealth affects young adults’ decisions to co-reside with their parents in Australia. As housing prices have risen and homeownership among younger cohorts has declined, the proportion of young adults living at home has increased significantly. Using data from the Household, Income and Labour Dynamics in Australia (HILDA) survey, this study investigates whether changes in parents’ self-reported housing wealth influence the likelihood of co-residence. A first-difference linear probability model is used to control for time-invariant factors, followed by an instrumental variable approach using local area median house prices to address measurement error. Results show that a 1% increase in perceived housing wealth is associated with a 0.0238 percentage point decrease in co-residence, with larger effects observed in the instrumental variable estimates. These findings suggest that housing wealth plays an important role in enabling young adults to move out and live independently. The study contributes to the literature on perceived wealth effects and intergenerational living arrangements, offering new evidence from the Australian context. It also highlights potential equity concerns, as access to housing wealth becomes increasingly important for life outcomes. This research has implications for housing policy and intergenerational inequality in Australia. |
| 2023 | Patrick Manning | Analyst - Reserve Bank of Australia | The Impact of US Housing Demand and Supply Shocks on The Australian Economy | This thesis develops an open economy structural vector autoregression (SVAR) model to determine how the Australian economy is affected by both a US housing demand shock and a US housing supply shock. Previous literature has focused on the how US housing shocks impact the local economy. Studies that focus on the international effects have either grouped Australia with other economies or have excluded Australia altogether, leaving a significant gap in literature of how the Australian economy is solely impacted. The results of the model indicate both a US housing demand and a US housing supply shock significantly impact the Australian economy, although in different ways. The most significant impact is that of the US house price shock upon Australian GDP which is large and persistent over time. The results contribute to the understanding of how Australian policymakers should incorporate the US housing market into policy decisions with this paper giving an example of how the key US housing variables could be incorporated into the Taylor rule. |
| 2023 | Serena McLennan | Economic Policy Assistant - Premier’s Department in Investment NSW | Gendered Narratives of HIV/AIDS in Malawi: a Text Analysis | Applying text-analysis techniques, I investigate gender-based differences in HIV/AIDS narratives among men and women in rural Malawi, assessing whether these change over the course of the epidemic. I utilise journals from the Malawi Journals Project, which capture narratives of everyday life in rural Malawi, as primary source material, and employ word clouds and sentiment analysis through dictionary methods to analyse variations in discourse and sentiment. My findings reveal distinctions along gender lines, which intensify in the later stages of the epidemic. Women consistently incorporate more familial language and express higher emotional content, aligning with anticipated gender differences, while men’s discourse focuses on disease transmission. However, an unexpected finding of prevailing positive sentiment towards HIV/AIDS is observed across all periods. Moreover, there is an increase divergence in sentiments by gender over time. These findings are suggestive of HIV/AIDS narratives responding to the interaction between the evolution of the epidemic and overarching social dynamics within Malawian society, which have resulted in women assuming a passive role in the epidemic. These findings suggest an opportunity for policymakers to utilise discourse and sentiment in developing HIV/AIDS policies that are responsive to diverse gender perspectives and needs. |
| 2023 | Genevieve Reich | Risk Analyst - Commonwealth Bank of Australia | Determining the impact of the 2004 Australian Baby Bonus on fertility rates using a synthetic control analysis | The persistent and longstanding issue of declining fertility in developed nations like Australia prompts a re-evaluation of the effectiveness of fertility-targeted policies. This thesis focuses on the 2004 Australian Baby Bonus (ABB), a prominent pronatalist policy that offered a $3,000 lump sum financial transfer for every child born after 1 July 2004. We employ the Synthetic Control Method, a data-driven approach, to construct a counterfactual scenario for Australia in the absence of the ABB’s introduction. Using aggregate country-level data from the World Bank World Development Indicators Database from 1998 to 2012, we carefully select suitable OECD countries as control units and incorporate relevant economic indicators as covariates. Our primary synthetic control model reveals a 6.82% increase in the total fertility rate induced by the ABB, surpassing prior estimates. Robustness tests, including alternative model specifications, consistently confirm a positive fertility effect. This comparative case study provides a nuanced and comprehensive analysis of the ABB’s impact, highlighting the importance of methodological choices in evaluating fertility-related policies. |
| 2023 | Georgia Rylah | Analyst - Reserve Bank of Australia | Beyond the Curve: Uncovering a Structural Break in the Japanese Beveridge Curve and Mismatch Dynamics Before the COVID-19 Pandemic | This paper studies whether the Japanese Beveridge Curve experienced a structural break prior to the COVID-19 pandemic. We utilise two levels of analysis to detect a break in the relationship between unemployment and vacancies and determine its timing and potential causes. We apply a Vector Error Correction Model to estimate the Japanese Beveridge Curve between January 2000 and June 2023. We detect a structural break in the Beveridge Curve in November 2019 and find evidence of change in the relationship between unemployment and vacancies as early as 2018. An extensive mismatch analysis uncovers that dynamics at the occupational and industrial level may be a potential contributing factor to the break. In particular, this study suggests that serviced-based industries and occupations may have contributed more to the break in the curve than others. The findings of this paper offer a new perspective on the recent dynamics of the Japanese labour market, bearing significant implications for policymakers. |
| 2023 | James Verschoor | Dealer Assistant - Income Asset Management Group | Measurement Error in the Trust-Risk Relationship: Examining the Relationship of Trust on Risk using an Obviously Related Instrumental Variables (ORIV) Approach | The link between risk on the decision to trust seems intuitive, one cannot know the true intentions of another person; it requires a leap of faith. Yet, when economists have examined this relationship, empirically they find no significant relationship. These results may be explained largely by a number of factors leading to measurement error. Using a trust game created by Berg Dickhaut and McCabe (1995) to measure trust and multiple measures of risk-preferences, this paper finds a positive and significant relationship of trust on risk. This paper adds to the literature by using an instrumental variables approach to account for measurement error known as Obviously Related Instrumental Variables (ORIV). In addition, this study adds to the literature examining the relationship of gender on trust, finding men display greater trust behaviours than women. |
| 2023 | Chengen Wu | TBA | Market Concentration and the Flattening Phillips Curve | This paper examines the flattening Phillips curve problem in macroeconomics. We review recent literature to compare past findings and identify gaps. The study then introduces a novel approach by integrating recent theories on the Phillips curve with market concentration. We develop an Autoregressive Distributed Lag (ARDL) New Keynesian Phillips Curve model, incorporating market concentration, to assess its impact on the Phillips curve. Empirical results indicate that market concentration does not statistically significantly influence the Phillips curve, suggesting that it neither contributes to its flattening nor strengthens the relationship. |
| 2022 | Andrew Evensen | Consultant - Compliance and Conduct, Regulatory and Compliance - KPMG | Lifetime Earnings in Australia over the Last Two Decades | This paper uses data from the HILDA Panel Survey to establish empirical facts on the evolution of lifetime and annual earnings and inequality between 2001 and 2020, for individuals between the ages of 25-39. Both men and women saw their lifetime earnings increase during this time across all three lifetime periods, with men seeing larger gains than women during all periods of the lifetime examined. From the first to the last cohort of men aged 25-39, lifetime earnings increased by 17%, whereas for women it grew by only 6%. Turning our attention to annual earnings, the opposite is the case. Women saw their earnings increase overall by 36.4% from 2001 to 2020—more significant earnings growth at the annual level than men, whose earnings only increased by 21.5%. At both the annual and lifetime level, earnings growth was significantly slowed during the 2010s, and in many cases was either stagnant or in fact decreased. Younger individuals were more affected by this trend, and men were affected more severely by this trend than women. In fact, we only see a notable divergence in earnings growth between the two genders after 2013. Despite this, the earnings of men increased more as they grew older, relative to women who consistently had similar levels of earnings across all age groups. Focusing now on inequality during this time, we see that it decreased slightly at both the lifetime and annual levels. By most metrics, inequality was higher for women than it was for men, and they had more pronounced differences in inequality between age groups. |
| 2022 | Jake Graceffo | Economist - NSW Treasury | The impact of COVID-19 stay-at-home orders on domestic violence assaults: A study of New South Wales | The impact of COVID-19 stay-at-home orders on domestic violence assaults: A study of New South Wales The introduction of stay-at-home orders (SAH) led to an immediate disruption of individuals’ daily routines with confinement to their households and changes in risk factors of domestic violence. Utilising data from the Bureau of Crime Statistics and Research (BOCSAR), I estimate the causal impact of SAH orders on domestic violence assaults in New South Wales (NSW). Adoption dates for these restrictions varied across LGAs, and as such, I implement a staggered difference-in-difference (DiD) specification. This model corrects for issues associated with weights in the presence of treatment effect heterogeneity across groups and time in staggered treatment settings. I find a significant decline in domestic violence assaults by 0.330 incidents per day of SAH orders, per 100,000. This corresponds to a 21% decline from pre-restriction daily incidence levels. These estimands are in contrast to various media articles and reports on summary data that outline increased domestic assaults. I discuss possible mechanisms that impact domestic violence with an analysis of these factors supporting empirical findings. |
| 2022 | Benaya Lie | Consultant - Ernst & Young | The effect and timing sensitivity of parenting style on cognitive and non-cognitive skills | This study explores the impact and timing sensitivity of parenting styles on children’s cognitive and non-cognitive skill development. While much policy attention is given to schools and neighborhoods, early development primarily occurs within the home, shaped by parental choices and behaviors. Using Australian panel data, this research examines how different parenting styles—authoritative, authoritarian, and permissive—affect children's outcomes over time. Applying a value-added model to address omitted variable bias, the study finds that the authoritative style, marked by high warmth and discipline, is optimal for both academic performance and non-cognitive traits like perseverance. The results reveal a notable timing dimension: for cognitive skills, early parenting (measured four years prior) has the strongest effect, whereas non-cognitive skills respond more to contemporaneous parenting. Australia’s cultural diversity provides a rich context for examining varied parenting approaches, and the findings support the theory of dynamic skill formation, where early investments yield compounding benefits. This research suggests that parenting style is not only influential but also time-sensitive, with implications for targeted, cost-effective interventions. Encouraging supportive and structured parenting practices may offer a powerful complement to traditional education policy in shaping long-term child development. |
| 2022 | Cassandra Russell | Policy Officer at Clean Energy Regulator | The Impact of Investments in Clean Electricity and Environmental Policy on Emissions for OECD Countries | This paper utilises an unbalanced panel of 38 OECD countries from 1980-2019 to investigate the relationship between investments in zero- and low-carbon sources of electricity capacity, GDP, CO2 and mitigation policy. Through Bai’s (2009) interactive fixed effects estimator accounting for cross sectional dependence, an interaction effect between capacity and GDP was explored. Investments or reductions in clean and dirty capacity per capita slightly altered the slope of the relationship between GDP per capita and CO2 per capita. However, the moderating effect of GDP per capita on the relationship between capacity variables per capita and CO2 per capita had a greater impact on slope. The interactions showed that simple increases in clean capacity and reductions in dirty capacity are not enough to change the direction of the relationship between GDP per capita and CO2 per capita. Rather, GDP per capita needs to be decarbonised deeper than just through the electricity sector. Using the two-stage instrumental variable (2SIV) and mean-group instrumental variable (MGIV) estimators of Cui, Norkute et al. (2022) to account for endogeneity, results were not as consistent. Further research needs to be done to account for endogeneity in exploring these interaction relationships while accounting for policy. However, when replacing capacity per capita variables with the clean share of total capacity, results were more consistent. Using both legislative acts and executive orders as instrumental variables, results from the MGIV estimator found that a one-unit increase in the share of clean capacity was associated with a 65.5% reduction in emissions. Similarly, a percentage increase in per capita energy consumption was associated with a 66.6% increase in CO2 per capita. |
| 2022 | Hayden Toohey | Economist - Oxford Economics Australia | Effectiveness of the Minimum Legal Smoking Age: Evidence from Australia | This thesis aims to test the effect of the mandatory minimum legal smoking age (MLSA) on smoking behaviours in Australian youth. Using a restricted access version of the Household Income and Labour Dynamics Survey in Australia (HILDA) that includes respondents’ date of birth, sex and self-reported smoking behaviours, I conduct a regression discontinuity design to estimate the causal effect of the MLSA on the proportion of youth who smoke and the average number of cigarettes consumed. I find no evidence that legal access to tobacco affects the proportion of youth who smoke or the average number of cigarettes they consume for my full sample. I find heterogeneous effects on the proportion of youth who smoke when subsetting the sample by sex, with significant effects found for males but not females, consistent with international literature on the effectiveness of legal tobacco access restrictions. However, the estimates for males are limited by sensitivity to bandwidth and polynomial specification, with statistically significant results not found in a subsequent robustness check. |
| 2021 | Tan Dat Huynh | PhD Student - Economics UTS | The Australian real exchange rate model and a Brainard-based (1967) policy recommendation | This thesis constructs a single equation error correction model to forecast the Australian dollar real trade weight index. It aims to confirm the long-run relationship between the exchange rate and the terms of trade through its in-sample explanatory power and out-of-sample forecast ability. It also determines which type of world interest rate should be used for the real interest rate differential variable, and, which particular type of commodity price sub-indices has the largest impact on the exchange rate. Furthermore, the paper also investigates the significance of the interest rate variable in the model to make a Brainard-based (1967) monetary policy recommendation. The main findings are that the error correction model outperforms random walk at various out-of-sample forecast horizons, the terms of trade and Australian-US yield gap form a long-run cointegrating relationship with the exchange rate, and the most suitable world interest rate to be used is the US interest rate. The error correction model specification is less robust when applying to fellow commodity currencies such as the Canadian and New Zealand Dollar. Finally, the model recommends the Reserve Bank of Australia should be less activist with their monetary policy. |
| 2021 | Erin Clarke | Research Assistant, ARTD Consultants | The 2018 Child Care Subsidy: Demand for Childcare and Maternal Labour Supply in Australia | I evaluate the effect of childcare subsidy reform on maternal labour supply and household’s childcare demand. Using HILDA survey data, I exploit a policy reform that reduced the cost of childcare for low-to middle-income households in Australia. Using a difference-in-differences design, I estimate the causal impact of the reform on household demand for childcare and the labour supply of mothers with preschool-aged children. Overall, I find significant positive effects of the reform on labour supply and childcare demand, with women in lower socio-economic groups seeing the strongest effects. My findings are in line with simple theoretical predictions and complement the existing empirical literature surrounding the relationship between maternal employment and childcare use. I also provide further insight into the labour supply and childcare demand decisions of Australian mothers with young children, which otherwise remains an unresolved issue. |
| 2021 | Scarlet Shannon | Support Analyst, Iress | Gender, Seniority, and the Dynamics of Economics Conferences | I study whether there are gender and seniority differences in presenters’ time use at a virtual economics conference. I conduct an empirical observational study to examine how the presenters use time in presentations, questions, and answers. I examine whether there are differences in terms of a range of conference session characteristics, such as field contribution and chair characteristics, and speaker characteristics, such as gender, seniority, co-authors, and citations. I compare how time is used amongst the speakers and whether the variables mentioned above influence the time used. I find evidence of seniority differences in presentation lengths and propensity for speakers to go over their time allocation and no gender differences in time use between speakers. I use the dataset and outcome variables to give an up-to-date insight into the current environment of economics and academic economic events. |
| 2021 | Rakesh Pillai | Business/Economics Graduate, Bayside Council (NSW) | The Impact of COVID-19 on Student Experience and Expectations Evidence through a Survey: Australia | To understand the impact of COVID-19 on university students experience in Australia, we replicate Aucejo, French, Araya, and Zafar's (2020) paper. We surveyed 152 students at the University of Technology Sydney. The study uses subjective treatment effects designed to uncover the causal impact of the pandemic on students current and expected outcomes. Overall, the results show adverse effects in the majority of dimensions. For example, 33% of students reported a negative casual effect in study hours, and 23% reported a negative causal effect in academic performance due to COVID-19. Results also demonstrate many areas of heterogeneity. For example, lower median income students make up almost 60% of students who expect to delay graduation and makeup 70% of students who changed majors due to COVID-19. These trends of casual impacts disproportionately hinder students from disadvantaged socioeconomic, low median income backgrounds, and students that lost jobs. With economic shocks produced by COVID-19, we explain distributional effects that occur among different groups. Finally, with the results found in this study and the results of Aucejo, French, Araya, and Zafar's (2020) paper, we distinguish common trends and disparities. |
| 2021 | Ben Robinson | Economic Policy Analyst, Commonwealth Treasury | An empirical study of crowding out in Australia using structural VAR models | This thesis aims to test the crowding out hypothesis by using structural vector autoregressive (SVAR) models to examine the relationship between government debt and interest rates in Australia between 1993 and 2021. Two government debt measures are used – the primary budget balance as a proportion of GDP, and the growth rate of the amount of securities in the market –and two interest rates are used –the interest rates on 3-year and 10-year government bonds. To fill the gaps identified in the literature, this study accounts for the role of monetary policy and Australia’s position as a relatively small, open economy. The results provide limited support for the crowding out hypothesis, and any positive relationship between government debt and interest rates is found to be smaller than those identified in previous studies. |
| 2021 | Alexander Koutalistras | Economic Analyst, Oxford Economics Australia | The Determinants of the Allais Paradox in the Loss Domain Under Expected Utility Theory | Since the inception of the Allais Paradox in 1953, there has been a profound effort to uncover its determinants. Is the Allais Paradox a result of the certainty effect, or is something else driving these deviations from expected utility theory? In this paper, I not only test the effectiveness of the certainty effect in explaining violations of expected utility in the loss domain but draw my attention to a more recent phenomenon proposed by Incekara-Hafalir et. al (2020), known as the zero effect. It states that perhaps, rather than an appeal to certainty, violations of expected utility theory are a result of an aversion to zero (Incekara-Hafalir and Stecher, 2012). Using data I obtain from an Amazon Mturk experiment, I adopt the novel Allais-style design proposed by Incekara-Hafalir et al. (2020), to determine which effect is more reasonable in outlining expected utility violations in the loss domain. I find that when I focus on Allais’ original, 2 common consequence design, the decisions made by participants are comparable to prior literature in terms of expected utility theory consistency. Conversely, when I consider patterns containing a variety of common consequences under Incekara-Hafalir et al.’s (2020) novel6-lottery design, I find much lower rates of expected utility consistency; possibly a result of an increase in the number of lotteries presented, errors, or inattention. If I adopt Allais’ original design but include one additional common consequence, I find strong support for the zero and certainty effects respectively, as well as the reverse certainty effect. Nonetheless, due to both the common consequence effect pattern and the reverse pattern to the common consequence effect being highly frequent, there is a strong possibility that participants made a high amount of random choices. |
| 2021 | David Xin Lu | Junior Settlements Analyst, Alinta Energy Australia | The effects of mental health differences on risk, time, and ambiguity preferences | Common mental illnesses such as depression and anxiety have a range of psychiatric and emotional symptoms that may affect how individuals make decisions. This study seeks test whether differences in mental health affect three key economic preferences: risk; time; and ambiguity, a type of risk preference. I utilise a newly collected twin dataset and implement a twin fixed effect model. The empirical strategy takes advantage of within-twin variation in mental health to provide a more suitable counterfactual group than previous studies. The estimated effects of mental illness on risk and ambiguity preference measures are small and statistically insignificant. While there are some statistically significant results for time preference measures that indicates that mental illness causes greater impatience and future bias, the results are not robust to multiple hypothesis testing and sensitive to different modelling of the independent variable. These results suggest that preferences are stable with mental health differences. |
