Kexin Gu, Ph.D. candidate in Finance
Bio
Kexin is a Ph.D. candidate in finance at Stevens Institute of Technology, expecting to graduate in 2025. She earned her master’s degree in statistics from Columbia University and completed a bachelor’s degree with a double major in mathematics and economics at Southwestern University of Finance and Economics.
Skillset
Her research interests include corporate innovation, behavioral finance, institutional investors and machine learning in finance. She is proficient in Python, R, Stata and MATLAB.
Dissertation Summary
Exploring Biases in Institutional Investment: Short-Term Innovation, AI in Financial Recommendations, and Cultural Factors
My research focuses on institutional investment strategies, AI in financial decision-making and behavioral biases in investment. I explore why investors favor short-term (exploitative) innovation over long-term (exploratory) strategies, how AI tools like ChatGPT may introduce biases in financial recommendations, and how cultural and behavioral factors affect investment decisions. Over the next ten years, I aim to develop more effective investment strategies that integrate cultural and behavioral insights, while addressing fairness concerns in AI-driven finance.
Innovation and Institutional Ownership: Exploitation versus Exploration. One key focus of my research hasbeen analyzing how institutional investors make decisions between firms that focus on short-term (exploitative) versus long-term (exploratory) innovation strategies. Exploitative innovation refers to companies that refine and optimizeexisting products or services togenerate immediate returns, while exploratory innovation involves riskier, long- term investments in new technologies or ideas. My research revealed that institutional investors overwhelmingly prefer firms that engage in exploitative innovation because these companies provide better short-term performance, which aligns with the investors’ focus on immediate returns. By analyzing patent- based data from 2003 to 2022, I demonstrated that portfolios with a higher concentration of exploitative firms consistently outperform portfolios that include more exploratory firms, especially for transient investors who prioritize short-term gains. This study highlighted the myopic behavior of institutional investors and showed a clear bias toward short-term profitability over long-term growth.
Race and Gender Biases in AI-Driven Financial Advice. Another focus of my current research examines Bias in AI-Generated Financial Advice: Investigating Gender and Racial Bias in Large Language Models (LLMs). With the rise of AI tools like ChatGPT in automating financial advice, there is growing concern that these models may introduce new forms of bias based on gender or race. While AI promises efficiency and the potential to reduce human biases, my research investigates whether these models—particularly in recommending mutual funds and other investment vehicles—exhibit bias in their outputs. My findings suggest that AI-driven tools may recommend fewer funds man- aged by black managers and more funds managed by male managers, pointing to potential racial bias embedded within the algorithms. This issue is critical, as it demonstrates that AI systems, often assumed to be neutral, can unintentionally perpetuate or even exacerbate historical biases found in human decision- making. Such bias can have serious implications in the financial industry, where investment decisions can lead to unequal access to opportunities and resources.
Cultural Bias and Fund Investment.Our research explores how cultural differences between fund managers and the companies they invest in affect investment decisions. We find that fund managers tend to invest more in companies where they feel a cultural connection to the leadership, even though this does not lead to better financial performance. Over time, managers become morecomfortable investing in companies they were initially culturally distant from. This pattern is most noticeable in largercompanies, smaller funds and when male managers are making the decisions.
This research shows that cultural bias plays a significant role in shaping investment strategies. Even in professionalsettings, fund managers may make decisions based on personal and cultural connections rather than focusing solely on financial factors. Our findings add to the understanding of how human biases, like in-group favoritism, influence investment behavior.
Academic Advisor
Sasha (Alexander) Rodivilov