Social Sciences

In-Koo Cho, PhD

PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, ECONOMICS

Collusion through Algorithms: Fact or Myth?

Economic activities are increasingly assisted by artificial intelligence, driven by an algorithm. As the algorithm becomes more powerful, one firm can closely monitor the competing firm's strategy in real time, which might facilitate collusion or anti-competitive agreement among firms. Observing seemingly anti-competitive behaviors by large high-tech companies, regulators, legal professionals, and economists raised suspicion of collusion through algorithms.

Existing economic papers on the collusion among algorithmic players examined oligopolistic Bertrand competitors producing strategic complements through numerical analysis. Generating the collusive outcomes numerically, the authors conclude that oligopolistic firms can learn to collude through algorithms.

We challenge the conventional reasoning by constructing a “counterexample.” Algorithmic duopolists can generate seemingly collusive behavior by precisely coordinating their actions to achieve the cartel price, even though we rigorously impose behavioral and institutional restrictions to block any explicit or implicit collusion. Roughly speaking, accusing the algorithmic firms of collusion in our example is essentially to blame the firms for processing the public market data through a well known statistical procedure.

A numerical analysis does not reveal the mechanism of how the algorithm processes the data to sustain the collusive outcomes but only shows the consequence of processing data. We aim to identify the features of an algorithm necessary for oligopolistic firms to learn to collude. Instead of numerical analysis, we opt for an analytic approach to investigate the price dynamics generated by interacting algorithms to understand how the collusive outcome arises, even though the firms do not have the ability or preference to collude.

Kiela Crabtree, PhD

ASSISTANT PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, POLITICAL SCIENCE

Missing, Murdered, Demobilized: The Impact of Childhood Violence on Electoral Outcomes

What is the legacy of discrimination and childhood violence on political participation? This project considers the political impact of the “Atlanta Child Murders” in Atlanta, Georgia, where the kidnapping and murder of 28 Black children between 1979 and 1981 revealed cross-cutting racial and class divisions within “Black Mecca.” Led by its first Black mayor, Maynard Jackson, the city seemed exemplary of Black political empowerment. Yet, a lack of urgency regarding the crisis – until it became a public relations problem – demonstrated disappointing limits on descriptive representation as well as leadership’s apathy toward lower socio-economic status constituents. The murders are imprinted in the collective memory of Black Atlantans, perhaps most of all among those who were themselves children at the time. This project unpacks the time period’s consequences by asking: How did the Murders impact Black electoral politics in the city? Was there a measurable impact on the later electoral engagement of those who, at the time, were children themselves? And how can such legacies best be identified? This study uses a multi-method design, combining original interviews and surveys with historical electoral returns to trace the political legacies of the Child Murders. Through a focus on Atlanta, this project draws the United States into a broader literature on the political legacies of violence around the world. In the process, I develop a causal framework for understanding the impact that violence in childhood has on political engagement later in life, expanding our understanding of the political legacies of violence.

Petra Creamer, PhD

ASSISTANT PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, MIDDLE EASTERN AND SOUTH ASIAN STUDIES

Rural Landscapes of Iron Age Imperial Mesopotamia

The Rural Landscapes of Iron Age Imperial Mesopotamia project (RLIIM) seeks funding to investigate the nature of the rural settlement Qach Rresh in the Assyrian Empire of the Iron Age Near East (c.900-600 BCE), located in modern-day Iraqi Kurdistan. This project involves targeted excavations and remote sensing of Qach Rresh and its surrounding landscape on the Erbil Plain, expanding from two seasons of excavation in 2022 and 2023 directed by the PI. Geophysical remote sensing investigations led by the PI at the site have revealed several structures potentially related to storage, administration, and associated pastoral practices. RLIIM seeks support from the URC for two seasons (6 weeks in June-July, 2 weeks in mid-August) of investigations at Qach Rresh and its surrounding area. Excavations will explore the Assyrian occupation phases of three large building complexes and their association with local agropastoral production and storage. Excavation will be accompanied by a vastly expanded remote sensing program using magnetic gradiometry to map site and landscape features. A robust sampling and analysis program (isotopic analysis of faunal and botanical remains; petrographic analysis of ceramic material; etc.) carried out by the PI and other core team members will investigate the extent of centralized administrative systems, adjustments in local subsistence strategies and human-environment interactions in premodern imperial spaces.

Caroline Fohlin, PhD

PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, ECONOMICS

Leveraging ‘Big Data’ and Artificial Intelligence to Study the Rise of Authoritarian Rule and the Economic Impacts of Political Turbulence in the 1920s and 30s

This research explores the economic causes and effects of the global decline of democracy and the rise of authoritarianism over the interwar period. With a primary emphasis on Germany, complemented by a comparative analysis of France and the United States, the research is the first to use truly “big data” and advanced artificial intelligence and natural language processing to investigate the interconnections between political regimes, social upheaval, financial systems, and the macroeconomy.

The current stage of the project focuses on data collection and language model training for variable creation and preliminary analysis. The project design and methods therefore involve obtaining source materials, ascertaining the most efficient approach collecting raw data, developing language models to identify and create the necessary variables, and specifying robust statistical models for hypothesis testing. To achieve these project milestones, the URC grant will fund undergraduate and graduate student research assistance in both computer science and economics, data collection and modelling costs, and sabbatical travel to work with colleagues and archival materials in Europe.

The funding will expand and accelerate the PI’s publication pipeline with articles that pose new questions and contribute new arguments to long-standing debates with the use of unique data sources. The work also contributes to the burgeoning research on applications of AI methods, language modelling, and the collection and analysis of “big data” in economics and financial history.

Allison LoPilato, PhD

ASSISTANT PROFESSOR, SCHOOL OF MEDICINE, PSYCHIATRY & BEHAVIORAL SCIENCES

Linking Decision-Making Alterations to Adolescent Suicide Risk

Adolescent suicide is an urgent public health crisis. In 2022, 22% of US adolescents reported seriously considering suicide, 18% made a specific plan, and 10% made an attempt. Despite decades of research, our current ability to predict suicide is only slightly better than chance. Thus, there is an urgent need to identify new and reliable predictors of suicide risk in adolescents. Recent work suggests that differences in how individuals make cost-benefit tradeoffs in decisions about pursuing reward and relief may be a promising marker of risk for suicidal thoughts and behaviors (STBs). However, more work is needed to determine which specific decision-making processes and tradeoffs are related to suicide risk in adolescents. The current project will determine whether alterations in decision-making processes related to reward, relief, effort, and delay underlie STBs in adolescents and whether these decision-making alterations are associated with childhood adversity. We will recruit a diverse sample of high-risk adolescents (ages 13-17 years; N=60) from outpatient clinics with: (a) recent suicide attempts with current ideation (n=20), (b) current suicide ideation with no attempt history (n=20), and (c) psychiatric controls with no STB history (n=20). Specific Aim 1 will characterize alterations in reward and relief decision-making processes in adolescents with STBs. Specific Aim 2 will examine whether decision-making alterations distinguish adolescents who attempt suicide from those who only think about it. Specific Aim 3 will identify whether decision-making alterations are associated with childhood adversity exposures. This line of work has the potential to significantly improve youth suicide prevention.

Stephen O'Connell, PhD

ASSOCIATE PROFESSOR, EMORY COLLEGE OF ARTS AND SCIENCES, ECONOMICS

From Displacement to Resilience: Aid, Economic Recovery, and Social Cohesion in Post-War Iraq

Civil conflicts leave profound and lasting impacts on individuals and communities affected by forced displacement. Exposure to violence, displacement, the loss of physical and human capital, and ruptured economic and social ties contribute to post-conflict environments exhibiting low levels of liquidity, trade, and capital investment as well as faltering trust and social cohesion. Livelihoods programming in post-conflict contexts has the potential to both support inclusion in labour markets and alleviate the root causes of low social cohesion by increasing economic opportunity, reducing inequalities and grievances, increasing contact, and lowering competition for services and resources. The role of livelihoods programs in improving social cohesion and securing durable solutions for displaced persons, however, is often assumed and only rarely studied. The International Organization for Migration (IOM) in Iraq operates a cash grant program in displacement-affected communities targeting poor individuals who have the potential to develop a business or enter paid employment. The programme provides several thousand new microentrepreneurs per year with \$2,000 each to establish a business– a value approximately two thirds of program participants' baseline annual income. In collaboration with IOM, we will randomize the selection of individual beneficiaries for the program to study the effects of this program on both economic and social well-being among applicants and their broader communities. In particular, the study will investigate how postwar recovery programming encourages sustainable livelihoods, economic resilience, and social cohesion in displacement-affected contexts.

Vilma Todri, PhD

ASSOCIATE PROFESSOR, GOIZUETA BUSINESS SCHOOL, INFORMATION SYSTEMS AND OPERATIONS MANAGEMENT

Ad-blocking Technologies and the Internet Economy

Despite the promising projections of digital economy, the rise of ad-blocking technologies poses a significant potential threat to the digital ecosystem and the primary business model that drives the Internet economy. To counteract the adverse effects of ad-blockers on the digital economy, online websites have started to employ anti-ad-blocking strategies, urging web visitors who use ad-blockers to disable them. This project aims to dissect the repercussions of anti-ad-blocking strategies employed by online publishers and their influence on the digital economy. Leveraging a difference-in-differences approach within a robust web-behavior panel dataset, we will explore user interactions with these strategies, assessing compliance and resistance behaviors against various factors, including website quality, firm size, and ad-load. For instance, users may reciprocate by complying with anti-ad-blocking requests for free website access, driven by the norm of reciprocity. Conversely, users might resist, perceiving these requests as impinging on their freedom and experiencing psychological reactance. This project aims to investigate the effectiveness of these increasingly prevalent anti-ad-blocking strategies and advance our knowledge of the digital economy’s sustainability, offering insights into the balance between user experience and economic viability. The project explores critical digital behavior patterns and stands to benefit society by informing the equilibrium between free content access and advertising revenue. This project makes significant contributions by addressing a critical gap in the literature and extending our understanding of digital consumer behavior and the strategic response of online publishers in the digital economy, yielding important practical implications for all parties of the online ecosystem.