Date of Award

Spring 6-7-2026

Document Type

Thesis (Undergraduate)

Department

Economics

First Advisor

Bruce Sacerdote

Abstract

Industrial organization and corporate finance papers that measure firm-level demand characteristics employ price, quantity, and product characteristic data that is unavailable in public financial disclosures. As a result, cross-sectional asset pricing literature attention to demand-side factors is limited. This paper constructs a demand shock measure from natural language processing of 28,263 publicly disclosed 10-K filings from 2001 to 2020. I find that a 1SD demand shock increases contemporaneously predicts a 1.8% increase (SE = 0.4%) in returns. Long-short portfolio sorted on demand shock generates 8.5% alpha (SE = 1.9%) against combined Fama-French and Q-factor models and a 0.91 sharpe ratio outperforming the 0.54 market. Examining firm characteristics as modulators of shock transmission reveals that a 1SD increase in short-term liquidity predicts a 2.9% (SE = 0.6%) amplification in the shock-return effect, while illiquid asset concentration and financial leverage dampen it. Liquidity amplification reflects partial overreaction to positive shocks, supporting literature tying liquidity to opaqueness of managerial effectiveness. Beyond mispricing, the liquidity amplification results suggest that liquidity is valued as a basket of real options for capacity expansion, the underlying project payoffs of which are sensitive to the state of demand.

Download

Share

COinS