Research
Job Market Paper
Task Efficiency and Signaling in the Age of GenAI: Effort Reallocation and Firm Value Effects
- Preliminary draft available upon request.
Abstract
In this paper, I study how Generative AI (GenAI) reshapes effort allocation and firm value by influencing the costs and signaling value of AI-assisted tasks. Using developer-level data from open-source projects linked to U.S. public firms and the launch of GitHub Copilot as a shock, I find that GenAI increases productivity in coding tasks but reduces the signaling value of such work for less-established developers. While senior developers benefit from increased efficiency in coding tasks, junior developers, whose contributions are less visible in an AI-assisted environment, create more-valuable projects as a more effective signal of ability. These changes in signaling incentives are reflected in selection of projects and languages, job mobility, promotion rates, and firm-level outcomes. Firms with more junior innovators exposed to AI see greater value creation from new projects, while non-innovative firms with senior teams capture efficiency gains. The findings shed new lights on the dual role of GenAI as both a productivity tool and a force reshaping labor market signaling.Presentations: AFA (2026, Scheduled), 10th European Conference on Corporate R&D and Innovation (2025, Scheduled), Zurich Workshop in AI+Economics (2025, Scheduled), Corporate Finance Day (2025), SGF Conference PhD Poster (2025), ERIM PhD Seminar (2024)
Awards: Corporate Finance Day Best PhD Paper Award
Working Papers
The Private Value of Open-Source Innovation, with Logan Emery and Chan Lim [SSRN]
Selected Presentations: FMA (Scheduled), GSU AI and FinTech Conference (Scheduled), Paris December Finance Meeting (Scheduled), ABFER (2025), CICF (2025), EFA (2025), ENTFIN (2025), Future Finance Fest (2025), Joint Conference with Allied KFA (2025), MFA (2025), SAIF Annual Research Conference (2025), University of Barcelona Micro Workshop (2025)
Abstract
Open-source innovation lacks the legal excludability viewed as essential for generating private value from innovation. Nonetheless, using investor reactions to GitHub releases by U.S. public firms from 2015-2023, we estimate an average private value of \$849,000 per project. Extrapolation to all projects during this period implies a total value of \$909 billion. Firms facing less competition release more projects, and both lower competition and restrictive licenses generate more private value. This value predicts firm growth, but peer benefits are modest. Overall, firms generate private value from open-source innovation by limiting spillovers, challenging the notion that open source fosters industry-wide growth.Campaign Rallies, Perceived Uncertainty, and Household Borrowing [SSRN]
Presentations: AFA PhD Poster Session (2025), EFiC Conference in Banking and Corporate Finance (2024), Augustin Cournot Doctoral Days (2024), ERIM PhD Seminar (2023)
Awards: ACDD2024 Best Paper Award
Abstract
This paper examines how political compaigns during the 2016 U.S. presidential election influences perceptions of economic uncertainty and subsequent household financial behaviors. Using a difference-in-differences approach, I show that Clinton's rallies reduced perceived economic uncertainty, particularly macro uncertainty. Moreover, areas hosting rallies showed an increase in P2P and mortgage loan applications after Clinton's visits, aligning with life-cycle models with precautionary motives. Effects are stronger in areas having higher initial level of economic uncertainty. In contrast, Trump's rallies did not significantly influence uncertainty perceptions or borrowing decisions. These findings shed light on a novel channel through which campaign information shapes real financial decisions, with effects contingent on the candidate involved.Publication
Boards of Banks, with Daniel Ferreira, Tom Kirchmaier, and Daniel Metzger [Link]
Forthcoming at the Journal of Corporate Finance