Social Sciences

Tucker Balch, PhD

PROFESSOR, GOIZUETA BUSINESS SCHOOL, FINANCE

Financial Market Manipulation by AI: Causes and Mitigation

We will investigate the agency problem in the context of AI-driven algorithmic trading, with a focus on how algorithms may inadvertently learn to engage in market manipulation such as spoofing. In finance, the agency problem arises when a “Principal” delegates tasks to an “Agent” whose interests may diverge. The agent is expected to act in the principal’s best interest—typically that of an investor or institution—however when the agent pursues actions aligned with its own objective, a conflict may arise at the expense of the principal. Unlike human agents, who follow objectives arising from incentives (e.g., bonuses) bounded by ethical guardrails, AI agents operate according to human-provided instructions and self-learning. Harmful strategies may arise even if the agent adheres to explicit human directions.

Our project will be among the first to examine this phenomenon in the context of finance market agency theory.  We will simulate a market including a set of trading agents following stereotypical strategies, as well as an AI trader using profit-maximizing strategies. Our goal is to determine the conditions under which the AI develops manipulative tactics without explicit programming. We will explore mitigation strategies, such as reward modifications, to reduce unethical behavior. Through agent-based modeling and simulation, we will also evaluate regulatory scenarios to assess their effectiveness.

We aim to inform ethical and regulatory frameworks to address the challenges posed by AI in financial markets. We request funding to support student researchers to implement the simulation infrastructure, computing resources for running simulations, and travel to engage collaborators.

Tetyana Balyuk, PhD

ASSISTANT PROFESSOR, GOIZUETA BUSINESS SCHOOL, FINANCE

Parental Income Inequality, Child Support, and Children's Consumption

Child support is one of the largest and most regular inter-household transfers, estimated at $30.6B annually in the U.S. However, these payments are largely unobservable due to the private nature of inter-household transactions, posing data challenges and creating gaps in the literature. Yet, child support is an important policy tool to re-allocate child-rearing expenses between parents. These payments (and non-payments) can affect the well-being of millions of children of divorced and never-married parents, especially when parents have low or substantially unequal incomes. Recent socioeconomic changes necessitate research on whether child support is effective in modern times. The effects of child support on parents’ saving/investing and borrowing choices, its effects on children’s consumption, and its heterogeneous effects are also understudied. This project will examine the impact of child support on parents’ consumption—saving decisions, financial choices, and child-rearing expenses, through robust empirical analyses using new micro data containing a comprehensive panel of consumer transactions and latest empirical methods with the aid of a newly developed theoretical framework. It will provide a set of stylized facts about parental income inequalities and child support payments; demonstrate how financing frictions, limited commitment, and modern custody arrangements affect the relation between child support and consumption—saving choices; measure marginal propensity to consume (MPC) with respect to child support; examine the causal effects of child support shocks; and estimate heterogeneous effects. The project is interdisciplinary in nature and can have impacts that inform policy and reach multiple disciplines, such as Economics, Sociology, Law, and Public Policy.

Leyla Eghbalzad, PhD

ASSISTANT PROFESSOR, OXFORD COLLEGE, PSYCHOLOGY

The Association Between Childhood Adversity and Statistical Learning Ability in Children: A Neuroimaging Study

Childhood adversity is critical factor influencing cognitive development, yet its impact on statistical learning (SL) ability remains underexplored. This study investigates the relationship between socioeconomic status (SES), childhood trauma, and SL—a cognitive process essential for language acquisition and broader cognitive functions. Existing research highlights that children from lower SES backgrounds experience reduced exposure to cognitive and linguistic stimulation, higher levels of environmental stress, and adverse childhood experiences, all of which may impair language and cognitive development. Moreover, SES disparities and early-life exposure to trauma manifest as structural and functional differences in brain regions associated with memory, language, and cognitive control. Building on these findings, this study proposes to examine how SES and childhood trauma interact to affect SL abilities in 30 children ages 6–12 years, a crucial developmental period. Using a combination of behavioral assessments and neuroimaging techniques, we will evaluate SL performance, language proficiency, and brain structure in diverse SES contexts. This research will advance our understanding of the neural mechanisms underlying adversity-related disparities in SL, contributing to developmental neuroscience and informing targeted interventions. By identifying how SES and trauma shape cognitive and neural processes, this study could add to our knowledge of the impact of childhood adversity with the ultimate goal of reducing educational inequities. Additionally, this project will provide undergraduate students at Oxford College with hands-on research opportunities, fostering their academic and professional growth while supporting broader institutional goals at Emory University.

Megan Mucioki, PhD

ASSISTANT PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, ENVIRONMENTAL SCIENCES

Elucidating cultural ecosystem services through relational health and wellbeing benefits with Indigenous Peoples in the United States: foundational planning

This proposal brings together health, environmental sciences, and Indigenous studies to challenge the way cultural ecosystem services (CES) are conceptualized and the way Indigenous health disparities are approached in the United States. In this proposal we lay the groundwork through planning efforts for research that will test the Indigenous Cultural Ecosystem Services theoretical model developed by PI Mucioki. The long-term goal of this project is to understand how environmental stewardship by Indigenous peoples, as one type of CES, garners quantifiable health and wellbeing benefits to Indigenous Peoples in the United States. This proposed research takes a community-based research and Indigenous methodologies approach, centering collaborating Indigenous communities as research partners. To successfully do so, it is imperative to have a phase of community discussion and planning to shape the research we will do together. The URC funds requested will be used to support the community planning phase prior to external grant funded research. The planning phase will allow new relationships to be formed and old ones to be strengthened with two communities, the Eastern Band of Cherokee and The Karuk Tribe, that have active forest co-stewardship agreements on public lands and are proposed collaborators for this work. Planning milestones include: one community discussion and five individual conversations during visits to each Tribal community, a two-day workshop at Emory with project collaborators to co-develop concepts of CES and health and our external research proposal, and the development of an external research proposal submitted to USDA-NIFA-AFRI Applied Science Program, Sustainable Agroecosystems program area.

Megan Reed, PhD

ASSISTANT PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, SOCIOLOGY

America's Chosen Families: Measuring the Prevalence and Implications of Voluntary Kinship in the United States

Voluntary kinship, also known as intentional families, fictive kinship, and chosen families, are the family-like relationships that people form outside of traditional definitions of kinship by marriage, blood, or adoption. The previous literature has mostly ignored voluntary kin and, when the have been studied, researchers have primarily employed qualitative case methods focusing only on specific populations like LGBTQ+, immigrant, and African American communities. The proposed project aims to provide the first nationally representative estimates of the prevalence of voluntary kinship in the United States. I will collect novel data on voluntary kin relationships using a new survey module that I designed for the nationally representative National Couples Health and Time Use Survey (NCHAT). This will allow me to be the first to document the prevalence and implications of voluntary kinship, focusing on how voluntary kinship practices, including the size of voluntary kin networks and support provided, vary across different groups in the United States. I will also study how voluntary kinship is associated with the mental health of LGBTQ+ individuals to test whether voluntary kin can buffer the negative effects of biological family estrangement and LGBTQ+ discrimination.

Holli Semetko, PhD

PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, POLITICAL SCIENCE

A Cross-National Comparative Study of the Drivers and Consequences of Affective Polarization in Mexico and India

The growing problem of affective political polarization in the U.S. is the focus of a substantial body of research but less is known about its prevalence in non-western multiparty contexts. To address this gap, we compare the drivers and consequences of affective political polarization in Mexico and India. We address a number of hypotheses framed as questions here: What are the drivers and consequences of affective polarization? Are more frequent and politically attentive users of social and traditional media more likely to be affective partisans than partisans or others including first-time voters? Are affective partisans stronger party supporters than partisans? Do affectively polarized partisans in winner and loser camps display more or less trust in government leaders? Do they display more extreme emotions? This cross-national comparative study will make important and novel contributions to the concept and measurement of negative partisan affect and the prevalence of affectively polarized partisans in two large understudied countries in the global south. Measuring attention to political news, media use, and more, the study models influences on vote choice among affectively polarized partisans as distinct from partisans for each party controlling for demographics. We also model influences on trust and emotional reactions among affective partisans, partisans, and others in winner and loser groups. Results will be published in academic journals and reported in news articles by the authors in both countries along with policy recommendations. Results will be the basis for external grant proposals to extend this research to a larger number of countries.