Applications of Textual Methods in Finance
This dissertation explores the nuanced dynamics of financial contracting on intermediation and firms’ human capital. Employing machine learning techniques, I unearth various dimensions of soft information embedded in the textual content of financial contracts. The first chapter expounds the margins of bank-firm bargaining relationship, while the subsequent chapters focus on the ex ante financial contract between workers and firms, as encapsulated by the text of job postings. In Chapter 1, “Corporate Liquidity under Basel III: The Credit Line Channel”, I investigate the impact of costly intermediation on bank lending responses and their contracting behavior. I do so under the setting of the Basel III Liquidity Coverage Ratio (LCR) rule, which imposed unprecedented liquidity requirements on banks. I show that the regulation curtails banks’ ability to originate credit lines, with banks seeking to pass on increased maintenance costs to borrowers. I introduce novel metrics drawn from a machine learning analysis of contractual agreements and demonstrate that banks retain greater control in credit lines. The result is a decline in credit line origination and a market that is unfavorable to borrowers. Financially unconstrained firms drive borrowing declines and turn to debt-financed cash for corporate liquidity, rendering them riskier. My results are novel in revealing changes to corporate liquidity preferences and risk profiles when intermediation is costly. In Chapter 2, “Corporate Hiring under Covid-19,” we study big data on job vacancy postings and unearth multiple facets of the impact of Covid-19 on corporate hiring. Firms disproportionately cut on hiring for high-skill positions, with financially constrained firms reducing skilled hiring the most. Applying machine learning methods to job-ad texts, we find that firms have skewed their hiring towards operationally-core functions. New positions take longer to fill, displaying greater flexibility regarding schedules and tasks. Financing constraints amplify pandemic-induced changes to the nature of positions firms seek to fill, with constrained firms’ new hires witnessing greater adjustments to jobs roles and employment arrangements. In Chapter 3, “Anti-Poaching Agreements, Corporate Hiring, and Innovation: Evidence from the Technology Industry,” we use the 2010 prosecution of U.S. technology firms engaging in anti-poaching agreements as a shock, and study the impact of labor market collusion on corporate hiring and innovation. During the collusive period, cartel firms displayed elevated job posting rates relative to comparable firms that were not party to these agreements. Occupation-level tests show that the effects were amplified in job roles critical to the firms’ operations. Textual analysis of job-ad descriptions provides evidence that cartel firms enjoyed greater bargaining power in the hiring process, with workers being offered lower flexibility, non-wage benefits, and training opportunities. Notably, cartel firms exhibited superior innovative capabilities over the collusive period, while the dissolution of the agreements led to a curtailment in their innovation output. Our results reveal important linkages between firms’ anti-competitive conduct in labor markets and their innovation and market valuations.