My research seeks to understand how beliefs, identity, expectations, and information environments shape financial and real estate behavior. Across projects, I study how people and firms act when information is incomplete, signals are noisy, and narratives or ideologies influence interpretation. Although the settings vary, from households to firms to neighborhoods, the central question is constant: When decision makers confront the same information, why do they reach different economic conclusions, and what does this mean for markets? I am interested in the moment when perception diverges from fundamentals, and in tracing how those divergences shape prices, choices, and spatial patterns over time. My work uses high-frequency market data, long-horizon real estate data, administrative voter files, corporate disclosure content, CEO turnover data, patent-level innovation measures, and spatial information systems to understand the mechanisms through which beliefs become economic outcomes.
What unifies the pipeline is a commitment to uncovering the hidden structure of human interpretation. Markets appear objective, yet they are built on judgment, identity, and expectations. My work shows that political alignment affects when families buy and sell homes, that neighborhood appreciation can be mistaken for infrastructure caused gains, that circulation and exposure patterns influence social sorting, that partisan identity shapes innovation in green energy, that analysts interpret the same macro environment differently based on ideology, and that CEOs nearing departure may use their final disclosure opportunity to shape valuations. At every level, I find that economic choices are filtered through cognitive and social lenses. My long-term goal is to map these filters, explain their consequences, and develop empirical strategies capable of distinguishing real causal effects from behavior that only appears causal.
My job-market paper, Political Alignment and Housing Transactions, establishes this agenda. I combine North Carolina voter registration data, MLS records, CoreLogic transactions, Case-Shiller volatility measures, and economic policy uncertainty indices to show that households who are politically aligned with the sitting president enter the housing market more often and transact more quickly. The result is a systematic difference in mobility, liquidity, and pricing across aligned and misaligned neighborhoods. This is the first paper to show how political sentiment affects real housing behavior, not only stated expectations.
My evidence on belief-driven selection appears again in When the Train Never Comes: Property Value Impacts from the Announcement and Cancellation of a Light Rail Project. Using repeat sales models and coarsened exact matching, my coauthors and I show that the neighborhoods selected for the proposed Durham–Orange light rail were already on an upward trajectory. Price increases predated the announcement, and the cancellation produced little to no measurable change after controlling for pretrends. This work clarifies that planners select areas already improving rather than causing improvement by selecting them. It deepens the theme that policy announcements are interpreted through prior expectations rather than producing automatic price reactions.
That same behavioral logic appears at the scale of social and political geography. The Mere Exposure Effect in Architecture demonstrates that repeated incidental contact within shared circulation paths increases preference and weakens segregation forces. Assessing the Geographic Sorting of Partisans in the United States extends this idea to voter movement. Using longitudinal voter records, the project asks whether households sort into aligned neighborhoods, whether they exit misaligned ones, and at which geographic scales sorting is most evident. Together, these papers explore how spatial patterns reflect accumulated psychological processes such as familiarity, identity, and perceived social fit.
At the firm level, Partisanship and Green Innovation: The Role of CEO Political Ideology in Environmental Investment tests whether partisan identity shapes long-horizon investment in environmental technologies. Using patent data, forward citations, CEO donations, and measures of board ideology, the project evaluates whether Democratic and Republican executives allocate resources differently to green innovation and whether national partisan cycles moderate these choices. Partisan Misalignment and Earnings Forecast: Evidence from Sell Side Analysts shows that analysts whose political identity is misaligned with the sitting president produce pessimistic forecasts, and that analyst competition attenuates this effect. Do CEOs Manipulate Stock Prices When They Know They Are Leaving? Evidence from Unexpected and Expected Turnover examines whether CEOs alter real activities, accruals, or disclosure tone when they anticipate leaving the firm, isolating CEO specific behavior from firm-level incentives.
Taken together, these projects form a coherent research pipeline: households, executives, analysts, and voters all interpret information through political, social, and spatial lenses. These lenses shape mobility, innovation, valuation, disclosure, and neighborhood composition. My work provides empirical strategies for distinguishing between actual economic effects and belief-driven reactions.
The next stage of my research extends this identity-centered framework to artificial intelligence. My proposal, Bias with a Price Tag: A Research Proposal on Systematic Human Bias in AI Financial Advice and Its Impact on Economic Outcomes, investigates whether identity-linked variation in AI-generated financial recommendations produces meaningful differences in long-run outcomes when individuals rely on those recommendations. Large language models draw their patterns from human-generated text, which means that the biases, heuristics, and interpretive structures present in human financial decision-making can reappear inside algorithmic systems even when the language seems neutral. This project tests identical financial profiles that differ only by attributes such as race, gender, partisanship, religion, or socioeconomic status and evaluates whether the AI responses create divergent paths for credit access, portfolio performance, wage negotiation, or retirement sufficiency. The goal is to determine when an alg
This project sits naturally at the end of my research pipeline because it examines a new environment in which interpretation is no longer limited to households, managers, analysts, or voters. It expands the agenda by exploring how identity-shaping information structures are transmitted into the tools that now mediate financial decision-making for millions of individuals. The project, therefore, brings the full pipeline together. It asks whether the forces that shape human perception also shape the recommendations produced by AI systems and whether those algorithmic interpretations lead to systematic differences in economic outcomes.