Designing Machine Learning Models to Predict Financial Distress in NYSE and NASDAQ Companies
Loading...
Issue Date
2026
Editor
Authors
Ramzan, Sana
License
Subject
Faculty of management
Abstract
The current dissertation by portfolio examines how machine learning models can be theoretically grounded to improve the early identification of financial distress and manipulation risk among firms listed on the NYSE and NASDAQ. Drawing on Merton’s strain theory as an integrative framework, the research conceptualizes financial distress not merely as an accounting outcome, but as a response to structural, market, and governance pressures that may motivate earnings manipulation. The dissertation consists of three interconnected portfolio papers. The first paper provides a systematic literature review of financial distress prediction research, identifying methodological, theoretical, and data-related gaps, particularly the dominance of financially driven models and the underutilization of governance and criminological perspectives. Building on these insights, the second paper empirically evaluates machine learning models that integrate financial, market and governance indicators with manipulation-linked signals, demonstrating how strain theory enhances interpretability beyond predictive accuracy. The third paper extends the framework by positioning corporate governance characteristics as early-warning signals of manipulation-linked distress, highlighting how governance weaknesses condition organizational responses to performance pressure. The synthesis paper integrates findings across the three portfolio papers, articulates their conceptual linkages, and demonstrates how the combined contributions advance theory, methodology, and practice. Collectively, the dissertation bridges crime and deviance theory with predictive analytics, offering an interdisciplinary and interpretable approach to financial distress prediction, while outlining pathways for scholarly research, regulatory relevance, and future commercialization of the research.
Description
2026