Research
Research
Text-based empirical projects in housing, political economy, and supply-chain risk.
Working Papers
Political Attention, Partisan Alignment, and Corporate Performance
Undergraduate Thesis, May 2024
I develop a “policy-taker” view of firms where managers respond to political contexts through disclosure rather than direct political ties. Using semi-supervised NLP on 220M+ earnings call passages and presidential documents, I construct text-based measures of Political Attention and Partisan Alignment that map corporate speech into policy themes. Linking these measures to firm outcomes shows that a one standard deviation increase in Political Attention raises ROA by roughly 2 percentage points, with lobbying expenditure mediating part of the effect and partisan alignment amplifying gains in politically exposed industries.
Work in Progress
The Numbers Game: Effects of Listing Price Formats in Housing Bargaining
with Haaris Mateen, Franklin Qian, Ye Zhang, and Tianxiang Zheng
We examine how round versus precise listing prices shape bargaining power in housing markets by combining 80M+ property listing records from a large proprietary multi-market housing transactions dataset with matched federally mandated mortgage reporting data and proprietary property transaction records. I lead the text-based measurement strategy that flags impatient or strategic sellers from listing remarks using Word2Vec embeddings, design the seller sorting index, and run heterogeneity analyses by neighborhood demographics and price segments. The goal is to open up the mechanism behind seemingly behavioral pricing rules with evidence grounded in text and administrative data.
Enterprise Supply Chain Risk and Network Status
with Prof. Yue Tang
We leverage 10,000+ Chinese annual reports to extract firm-year risk language and overlay it onto a directed supply-chain network for all A-share firms. My focus is on fusing the text-based risk measure with network centrality to trace how position in the production web insulates firms from external shocks, especially during policy or trade disruptions.
Research Experience
Franklin Qian, UNC Kenan-Flagler · Mar 2025 – present
Design seller sorting measures, build Word2Vec-based impatience scores for property listing remarks, and lead the matching of federally mandated mortgage reporting data with proprietary property transaction records to evaluate demographic heterogeneity.
Letian Zhang, Harvard Business School / Northwestern Kellogg · Aug 2024 – Mar 2025
Built longitudinal database of 600+ political leaders with coded ideology, constraints, and transitions; assembled a multilingual corpus of 9,000+ political speeches for future text analysis on corporate human-capital responses.
Yue Tang, Southwestern University of Finance and Economics · Feb 2023 – Jan 2024
Co-authored the initial draft of a supply-chain risk paper by creating a Chinese-language text-mining pipeline and estimating fixed-effects models tying network centrality to risk exposure.