My research interests are in asset pricing, real estate finance, and household finance. My work also explores innovation and political economy, applications of artificial intelligence and natural language processing in finance, and the implications of AI for financial markets.
Before joining UBC, I was a predoctoral research fellow at Cornell University. I hold an M.A. from the University of Chicago, an M.Sc. from the London School of Economics, and a B.Phil from Xiamen University.
Does the market price technological spillovers? If so, what determines the spillover value of innovation? I construct a patent-level measure of innovation’s spillover value from peer firms’ equity reactions to U.S. patent grants from 1926 to 2023. Around patent grants, peer firms in the patenting firms’ industries earn higher three-day market-adjusted returns than firms outside the affected industries. Returns also increase with firms’ technological similarity to the newly granted patents. The absolute magnitude of spillover value predicts forward citations. An additional dollar of private patent value is associated with approximately $0.20 of net equity value among publicly traded industry peers. Peer firms’ spillover exposure predicts long-run output and employment growth conditional. Spillover effects on growth are stronger in more competitive industries. Higher spillover exposure also predicts faster growth in investment and acquisition expenditure, as well as increased inventive activity. Spillover pass-through varies with innovation characteristics, industry conditions, and the information environment. It is larger for breakthrough patents, process patents, and technologies in rapidly evolving fields, and weaker in industries with greater market power or tighter financial constraints. Measured pass-through is also stronger in larger industries and for patents issued by more actively traded firms, consistent with investor attention facilitating the recognition of external innovation opportunities in prices.
We examine whether partisanship shapes home purchase decisions and their financial consequences using a dataset that links individual housing transactions to voter registration records in the United States from 2010 to 2023. We show that Republicans are nearly twice as likely as Democrats to purchase homes and realize 0.69 percentage point higher housing returns. Individuals are also significantly more likely to purchase a home when their affiliated party controls the White House, an effect we identify using the surprise 2016 election as a shock. The alignment effect is strongest for investment-driven purchases and much weaker for consumption-driven purchases. Using survey data on housing market expectations, we provide direct evidence for the mechanism: partisan alignment raises optimism about housing market conditions, and this expectations channel drives purchase decisions. Partisan alignment also distorts purchase timing: aligned buyers tend to purchase too early in a new administration, when partisan optimism is strongest, which is associated with lower subsequent returns. These individual-level shifts in housing demand aggregate to local housing markets, producing higher house prices but lower subsequent housing returns, a pattern consistent with belief-driven overvaluation.
Chronologically Consistent Large Language Models, Value of Patent, and Entrepreneurship Success, Draft Coming Soon
Research overview
This project asks whether patent text can predict inventions' market value and entrepreneurial success using language models designed to mitigate look-ahead bias.
Institution Ownership Concentration and Payout
Using institutional ownership 13F data from 1980 to 2020, this paper examines the relationship between firms' payout policy and ownership concentration. Consistent with the existence of conflicts of interests between large concentrated institutional shareholders and management, I find that firms that experience an increase in ownership by large and more concentrated institutional shareholders experience a decrease in payout. The result holds both in the panel regressions with firm fixed effects and industry-by-year fixed effects and in a difference-in-differences design based on significant increases in concentrated institutional ownership. The result is robust to alternative payout measures in both the panel and DID specifications. The result is concentrated in industries with high concentration and firms with staggered board structures, which prior literature suggests have bigger agency problems. The evidence suggests that institutional ownership concentration mitigates the concern of agency problems between large concentrated institutional shareholders and management and reduces the need to force cash payout.
2022 UBC Sauder First-year Paper.
Selected Work in Progress
STEM Narrative and Asset Price
How do narratives about technological progress shape the cross-section of expected stock returns? Using rolling Word2Vec embeddings trained on a century of Wall Street Journal articles, this paper constructs a Science, Technology, Engineering, and Mathematics (STEM) Narrative Index that captures the time-varying salience of technological discourse. Across 138 anomaly portfolios, greater positive exposure to the STEM Narrative factor is associated with higher expected returns. The factor provides incremental explanatory power beyond the five-factor model of Fama and French (2015) and the augmented q-factor model of Hou et al. (2021), suggesting that technological narratives capture a dimension of expected returns not fully explained by standard asset pricing models.
AI Capitalists: Artificial Intelligence and the Cost of Capital
This paper examines how exposure to artificial intelligence (AI) risk relates to firms’ cost of capital. I construct a firm-level measure of AI risk exposure by combining newly collected data on AI attention in the New York Times from 2004 to 2023 with stock return comovement. AI exposure is positively associated with market beta and negatively associated with idiosyncratic volatility. Firms with greater AI exposure have a lower cost of capital, measured using both firms’ self-reported perceptions and option-implied estimates. A one-standard-deviation increase in AI exposure is associated with a 6–15 basis point reduction in the annual cost of capital, consistent with a negative price of AI risk. Additional evidence suggests that concerns about labor displacement contribute to hedging demand for AI-exposed stocks.
Economists' Incentives, Outside Consulting, and the Production of Economic Knowledge
How does consulting shape economists’ research productivity and agendas? Using a novel dataset linking economists’ publication records to their CVs, this paper examines how consulting affects both the volume and scope of academic research. To identify causal effects, the empirical strategy uses a Bartik-style instrumental variable design that exploits shifts in industry and academic labor demand to isolate variation in consulting participation. The findings reveal contrasting patterns across fields: consulting among finance economists is associated with greater publication output and broader research agendas, whereas consulting among industrial organization economists is associated with fewer publications and narrower research agendas. These results suggest that engagement with industry can complement or displace academic research, with implications for both scientific productivity and the direction of knowledge production.
Teaching Experience
University of British Columbia
COMM 370 / BAFI 502Corporate Finance
COMM 371Investment Theory
COMM 374Applied Financial Markets
COMM 375Behavioral Finance
COMM 471 / BAFI 507Mergers and Acquisitions
University of Chicago
ECON 3214Computational Methods in Economics
ECON 2020The Elements of Economic Analysis III (Intermediate Macroeconomics)