https://hdl.handle.net/1721.1/138916 2026-04-06T11:47:23Z https://hdl.handle.net/1721.1/150502 From ELIZA to ChatGPT: The Evolution of NLP and Financial Applications Lo, Andrew W; Singh, Manish; ChatGPT Natural language processing (NLP) has revolutionized the financial industry, providing advanced techniques for the processing, analyzing, and understanding of unstructured financial text. We provide a comprehensive overview of the historical development of NLP, starting from early rules-based approaches to recent advances in deep-learning-based NLP models. We also discuss applications of NLP in finance along with its challenges, including data scarcity and adversarial examples, and speculate about the future of NLP in the financial industry. To illustrate the capability of current NLP models, we employ a state-of-the-art chatbot as a co-author of this article. KEY FINDINGS (1) The use of NLP in finance has evolved significantly over the past few decades with the growth of data, storage, and computational power. NLP is now used for wide range of sophisticated tasks including asset management, risk management, and impact investing. (2) The development of deep-learning-based large-language models like GPT-3/ChatGPT have significantly advanced the applications of NLP in finance. These models have the ability to understand and generate human-like language, making them useful for various tasks, including assisting in the writing of this article. (3) Solving problems related to data bias, high computational needs, and inaccurate responses generated by the models will make NLP models even more accessible and indispensable. 2023-04-20T00:00:00Z https://hdl.handle.net/1721.1/143562 The Effects of Spending Rules and Asset Allocation on Non-Profit Endowments Halem, Zachery M.; Lo, Andrew W; Matveyev, Egor; Quraishi, Sarah The long-run impact and implications of an endowment’s spending policy and asset allocation decisions are examined. Using a dynamic model, the authors explore how different endowment spending rules influence the dynamics of an endowment’s size and future spending. They find that different parameters within each spending rule have significant long-term impact on wealth accumulation and spending capacity. Using Merton's (1993) endowment model and compiled asset allocation data, they estimate the intertemporal preferences and risk aversion of several major endowments and find significant variation across endowments in their propensity to increase portfolio risk in response to increased spending needs. Research support from the MIT Laboratory for Financial Engineering is gratefully acknowledged. We thank the MIT Investment Management Company, particularly Seth Alexander and Tom Wieand, and the editor for helpful comments and discussion, and Jayna Cummings for editorial support. The views and opinions expressed in this paper are those of the authors only, and do not necessarily represent the views and opinions of any institution or agency, any of their affiliates or employees, or any of the individuals acknowledged above. 2022-06-27T00:00:00Z https://hdl.handle.net/1721.1/143561 An Artificial Intelligence-Based Industry Peer Grouping System Bonne, George; Lo, Andrew W; Prabhakaran, Abilash; Siah, Kien Wei; Singh, Manish; Wang, Xinxin; Zangari, Peter; Zhang, Howard In this work, the authors develop a data-driven peer grouping system using artificial intelligence (AI) tools to capture market perception and, in turn, group companies into clusters at various levels of granularity. In addition, they develop a continuous measure of similarity between companies; use this measure to group companies into clusters and construct hedged portfolios. In the peer groupings, companies grouped in the same clusters had strong homogeneous risk and return profiles, while different clusters of companies had diverse, varying risk exposures. The authors extensively evaluated the clusters and found that companies grouped together by their method had higher out-of-sample return correlation but lower stability and interpretability than companies grouped by a standard industry classification system. The authors also develop an interactive visualization system for identifying AI-based clusters and similar companies. We benefited from the methodology of MSCI Peer Similarity Scores product and the data and computing support from the MSCI Data Science Platform. We thank Roman Kouzmenko and Manuel Rueda for insightful discussions. The views and opinions expressed in this article are those of the authors only, and do not necessarily represent the views and opinions of any institution or agency, any of their affiliates or employees, or any of the individuals acknowledged above. 2022-01-01T00:00:00Z https://hdl.handle.net/1721.1/141712 When Do Investors Freak Out? Machine Learning Predictions of Panic Selling Elkind, Daniel; Kaminski, Kathryn; Lo, Andrew W.; Siah, Kien Wei; Wong, Chi Heem Using a novel dataset of 653,455 individual brokerage accounts belonging to 298,556 households, we document the frequency, timing, and duration of panic sales, which we define as a decline of 90% of a household account’s equity assets over the course of one month, of which 50% or more is due to trades. We find that a disproportionate number of households make panic sales when there are sharp market downturns, a phenomenon we call ‘freaking out.’ We show that panic selling and freak-outs are predictable and fundamentally different from other well-known behavioral patterns such as overtrading or the disposition effect. 2022-01-01T00:00:00Z