AI in Proxy Voting: Recommendation Analysis and Prediction

• Applicant: Professor Dr. Stefan Ruenzi
• Subject Area: Accounting and Finance
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 575781215

Institutional investors, such as mutual funds, regularly cast votes on corporate matters. These decisions are often guided by the advice of proxy advisors like Institutional Shareholder Services (ISS), whose recommendations significantly influence voting outcomes and corporate governance. However, the recommendation process is opaque, potential conflicts of interest are underexplored, and private (retail) investors are excluded from this information, leading to structural information asymmetries and reduced retail participation. This project addresses two major goals: (1) Evaluating AI Models as Automated Voting Advisors: Can artificial intelligence reliably interpret voting guidelines and produce sound recommendations? We test various AI models to understand their strengths and limitations in processing complex governance proposals. Through careful comparison with both human expert assessments and professional recommendations, we identify where AI excels and where it falls short. This work not only contributes to our understanding of AI in financial contexts but also leads to a practical tool that could give retail investors access to automated voting guidance previously available only to institutions. (2) Investigating Conflicts of Interest in Proxy Voting Advice: We systematically compare AI-generated recommendations with actual recommendations issued by ISS to identify patterns of deviation. We examine whether these deviations are associated with potential conflicts of interest, such as when shareholders submitting proposals are ISS clients. Our findings will inform ongoing academic and regulatory debates on the transparency and accountability of proxy advisors. By combining AI innovation, empirical financial research, and regulatory relevance, the project offers valuable insights for scholars, market participants, and policymakers alike. The expected outcomes promote fairness and transparency in corporate decision-making while demonstrating a novel application of AI in financial research and public engagement.

• DFG Programme: Research Grants