18693 research outputs found
Sort by
Migration Fear, Race-Related Biases and Analyst Forecast Accuracy
No full textThis study examines whether heightened migration fear impairs minority analysts\u27 ability to access and convey accurate financial information. Using aggregate U.S. newspaper article counts regarding migration fear as a proxy for society wide sentiment about ethnic minorities and a sample of 1.3 million quarterly earnings announcements over 1990-2023, we find that elevated levels of the Migration Fear Index increase absolute forecast errors for Non-White analysts. For the average firm in our sample, a one standard deviation increase in migration fear leads to an additional 4 cents in EPS forecast error for Non-White analysts. Our results are robust across alternative measures of xenophobia, analyst accuracy, entropy-balanced samples, and fixed effects models. Using shocks to migration fear from the launch of the Make America Great Again Campaign in 2015, we observe that the shock selectively impairs Non-White analysts\u27 forecast accuracy. Further analyses reveal that migration fear\u27s impact on Non-White analyst forecast accuracy is driven by ethnic out-group bias, as the negative effects of xenophobia on Non-White analysts disappear when the CEO is also Non-White or when the firm is headquartered in a pro-immigrant area. Findings suggest that elevated levels of the Migration Fear Index lead to reduced information sharing with Non-White analysts, limiting their access to private information from management and thereby increasing their forecast error. We find no evidence of migration fear reducing Non-White accuracy through impaired cognitive processing capacity. These findings underscore the role of societal biases in shaping information access across capital markets, potentially hindering both price discovery and minority analysts\u27 career development
Cross Trading in the Corporate Bond Market
No full textWe examine cross-trading by mutual funds in corporate bonds. Because internally matched trades are not observable, we construct two measures that rely on reported trade volume and opposite-signed trades within a family. We find that cross-trading is common--more than 5% of bond-family-quarters have positive indicators. There is large variation across families and higher crossing activity for illiquid and hard-to-obtain bonds. Cross-trading is particularly elevated around demand shocks (e.g., maturity cutoffs and credit rating changes), indicating that cross-trading is beneficial in times of stress. We document large transaction cost savings, although these savings have diminished following a new regulation that significantly limits cross trading in corporate bonds
Evaluating the Implementation and Impact of Illinois’ Pretrial Fairness Act
No full textOn September 18, 2023, Illinois implemented the Pretrial Fairness Act (PFA). This landmark reform eliminated cash bail and introduced sweeping changes to the state’s pretrial system. The PFA establishes new release and detention factors, restricts pretrial detention, and limits post-release modification of release conditions. Professors David Olson and Don Stemen of Loyola University’s Center for Criminal Justice will share findings from their ongoing, long-term evaluation of the PFA’s implementation. They will also highlight variations in the PFA’s implementation and impact across urban and rural Illinois jurisdictions
AI-Powered Compliance: Accelerating efficiency and decision-making for Compliance related inquiries.
No full textThis research examines the potential of an AI-powered chatbot to streamline compliance workflows by reducing the time and effort required to locate and interpret complex compliance documents. The prototype integrates a centralized MySQL-based document repository, a contextual document querying engine, and a Streamlit web interface, enabling employees to retrieve accurate, document-backed answers within seconds. The system supports both stored and user-uploaded documents, with features such as automated summarization and source citations to enhance transparency and trust. Manual evaluation demonstrated notable gains in efficiency and accuracy compared to traditional search methods, with strong potential to improve adherence to compliance policies. Future work will focus on scaling document coverage, implementing automated performance testing, and strengthening ethical safeguards, including privacy protections and explainability features
3D Multi-Threaded AI Navigation with Pathfinding and Obstacle Avoidance
No full textIn this thesis, I developed a 3D multi-threaded AI navigation system using my own custom-built C++ game engine. The system combines triangle-based A* pathfinding with real-time obstacle avoidance using a set of velocity-obstacle algorithms. It is designed to support large numbers of agents navigating complex environments while avoiding collisions. I created two main simulation modes: Navigation Mode, which integrates A* with ORCA to handle large-scale pathfinding and movement, and Obstacle Avoidance Mode, which allows direct comparison between VO, RVO, HRVO, and ORCA in a controlled test setting.
The terrain is procedurally generated using Perlin noise, and this terrain data is then used to generate a NavMesh composed of walkable triangles. Each triangle connects to its neighbors through shared edges, allowing agents to move fluidly across the mesh. My A* algorithm operates directly on this triangle structure and includes a pruning step that removes unnecessary waypoints to produce cleaner, more efficient paths. Pathfinding is fully multi-threaded, with each request running as an independent job using a custom job system, enabling fast and scalable computation.
For obstacle avoidance, I implemented several velocity-obstacle-based algorithms that allow agents to dynamically adjust their movement based on nearby agents. ORCA provides the most advanced solution, calculating half-plane constraints to ensure smooth and collision-free navigation. To support debugging and evaluation, I implemented visualizations for velocity cones, direction adjustments, and constraint regions.
Agents use physics-based movement with smooth acceleration, turning, and terrain-aware positioning. In Obstacle Avoidance Mode, agents are placed in a sandbox environment loaded from an OBJ model, which creates a closed space for consistent behavior testing. This project gave me the opportunity to explore how pathfinding, real-time avoidance, and multi-threaded systems can be combined to create scalable and intelligent 3D AI navigation. It demonstrates how custom-built solutions can meet the demands of modern simulation and game environments
The Implications of Resale Channels in the Apparel Industry
No full textThere is widespread perception that resale platforms that allow consumers to buy and sell used apparel play an important role in reducing production and waste. Many consumers possess clothes that they rarely use because their individual utility has eroded, due to changes in fit or personal tastes. Nearly all existing resale channels, both those operated by branded apparel manufacturers and those that operate independently, promote themselves as helping to get clothes out of closets or landfills by finding new homes for them. While resale channels seem well-positioned to transfer products from consumers who no longer value them highly to those who do, whether they will necessarily benefit branded apparel manufacturers, or reduce the quantity of new units produced are open questions. We develop a game-theoretic infinite-horizon model in which a brand chooses production and pricing while interacting with resale channels. We first consider brand-operated resale and show that resale increases the quantity of new units only when it is also socially efficient. We then consider the impact of a third-party resale channel on a brand that does not operate its own. In this case, the brand never benefits from resale, and the possibility of resale increases output of new units for a range of intermediate transaction costs for which this is not socially efficient. Finally, we extend the analysis to competitive resale markets. While competition can replicate brand-controlled outcomes, it can also exacerbate overproduction for a range of intermediate transaction costs. Our results demonstrate that resale channels are not uniformly beneficial, especially when operated by third-parties. Without well-structured relationships between brands and resale platforms or clear policies about which products a resale platform will trade, resale may encourage overproduction. This insight should inform decisions made by brands, resale platforms, and regulators
Investigating Differences in Cognitive Function Between Middle to Older Age Adults With and Without Asthma
No full textAsthma, age, and overall inflammation have been implicated in cognitive decline. This study examined the relationship between asthma, age, and systemic inflammatory cell activation with cognitive performance among middle-aged and older adults. Participants (N = 132; 72 asthma, 60 controls) completed neuropsychological tests assessing executive control, episodic memory, and processing speed. Blood samples were analyzed for neutrophil–lymphocyte ratio (NLR) and absolute eosinophil count as markers of inflammation. Multivariate multilevel models tested whether higher inflammation predicted poorer cognition, and whether these associations differed by asthma status. Results indicate that neither inflammatory marker were significantly associated with cognitive performance for either group. Participants with asthma demonstrated better overall cognitive performance than controls, particularly in episodic memory. Additionally, older age was found to be associated with poorer cognitive performance overall, however in the control group age was more strangely associated with poor performance than the asthma group. Within the asthma subgroup respiratory medications were not significantly related to cognition. Findings suggest that in a sample with inflammation levels within normal ranges, systemic inflammatory cell activation is not associated with cognitive performance. Enhanced memory performance among asthma participants may reflect indirect benefits of disease management or pro-health behaviors. Future work should aim to include samples with more diverse asthma manifestations, additional inflammatory biomarkers (e.g., CRP, cytokines), and broader socioeconomic representation to better clarify mechanisms linking inflammation, asthma, and cognitive aging
Willful Ignorance or Embracing AI to Find Prior Art? USPTO Misses the Mark
No full textArtificial intelligence (AI) is frequently in the news. Policymakers and business executives must decide whether it is a revolutionary phenomenon, likely to overturn established enterprises and practices, or if it is a mere modest advancement in decades-old natural language capability over-advertised by Silicon Valley seeking the “next big thing” and hyped by reporters eager for the next byline. The patent office and the patent bar are not immune from the turmoil. The USPTO has issued several policies and inquiries related to the impact of AI on various aspects of patent prosecution. Most recent is a Request for Comments on the impact of AI on patentability determinations, in particular its impact on discovering and disclosing prior art. Until mid-2025, the USPTO had been regrettably defensive in its public statements on the subject. The patent bar, however, in its comments to the USPTO, recognizes the potential of AI technology. Generative AI can bring the patent system closer to achieving its goal of being an engine of innovation and improving public access to inventions. It has the capability to make patent prosecution more ac- curate and more efficient by assessing whether a new application for a patent is anticipated by older references and whether it is obvious to a person of ordinary skill using AI search technology. Examples show the power of generative AI in this context and encourage the patent office to embrace it, rather than keeping its distance