Can ChatGPT improve investment decisions? From a portfolio management perspective

Hyungjin Ko; Jaewook Lee

doi:10.1016/j.frl.2024.105433

Can ChatGPT improve investment decisions? From a portfolio management perspective

Hyungjin Ko, Jaewook Lee

Seoul National University

Research output: Contribution to journal › Article › peer-review

41 Scopus citations

Abstract

We examine ChatGPT, a prominent Large Language Model (LLM), in supporting portfolio management with a focus on asset selection and diversification through quantitative methods. We use ChatGPT to select assets from various asset classes and evaluate the diversification effect of its selections. Our results suggest that ChatGPT's selections are statistically significantly better in diversity index than randomly selected assets. We also construct portfolios based on ChatGPT's selections and find that they outperform portfolios built on randomly selected assets. Overall, our study contributes to a better understanding of the role of LLMs like ChatGPT as potential assistants for portfolio managers.

Original language	English
Article number	105433
Journal	Finance Research Letters
Volume	64
DOIs	https://doi.org/10.1016/j.frl.2024.105433
State	Published - Jun 2024
Externally published	Yes

Keywords

Asset selection
ChatGPT
Diversification effect
FinTech
Investment
Large Language Model
LLM
Portfolio management

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.frl.2024.105433

Cite this

@article{7ed03b3d0f9245ea87501df86ef24e02,

title = "Can ChatGPT improve investment decisions? From a portfolio management perspective",

abstract = "We examine ChatGPT, a prominent Large Language Model (LLM), in supporting portfolio management with a focus on asset selection and diversification through quantitative methods. We use ChatGPT to select assets from various asset classes and evaluate the diversification effect of its selections. Our results suggest that ChatGPT's selections are statistically significantly better in diversity index than randomly selected assets. We also construct portfolios based on ChatGPT's selections and find that they outperform portfolios built on randomly selected assets. Overall, our study contributes to a better understanding of the role of LLMs like ChatGPT as potential assistants for portfolio managers.",

keywords = "Asset selection, ChatGPT, Diversification effect, FinTech, Investment, Large Language Model, LLM, Portfolio management",

author = "Hyungjin Ko and Jaewook Lee",

note = "Publisher Copyright: {\textcopyright} 2024 Elsevier Inc.",

year = "2024",

month = jun,

doi = "10.1016/j.frl.2024.105433",

language = "English",

volume = "64",

journal = "Finance Research Letters",

issn = "1544-6123",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - Can ChatGPT improve investment decisions? From a portfolio management perspective

AU - Ko, Hyungjin

AU - Lee, Jaewook

PY - 2024/6

Y1 - 2024/6

N2 - We examine ChatGPT, a prominent Large Language Model (LLM), in supporting portfolio management with a focus on asset selection and diversification through quantitative methods. We use ChatGPT to select assets from various asset classes and evaluate the diversification effect of its selections. Our results suggest that ChatGPT's selections are statistically significantly better in diversity index than randomly selected assets. We also construct portfolios based on ChatGPT's selections and find that they outperform portfolios built on randomly selected assets. Overall, our study contributes to a better understanding of the role of LLMs like ChatGPT as potential assistants for portfolio managers.

AB - We examine ChatGPT, a prominent Large Language Model (LLM), in supporting portfolio management with a focus on asset selection and diversification through quantitative methods. We use ChatGPT to select assets from various asset classes and evaluate the diversification effect of its selections. Our results suggest that ChatGPT's selections are statistically significantly better in diversity index than randomly selected assets. We also construct portfolios based on ChatGPT's selections and find that they outperform portfolios built on randomly selected assets. Overall, our study contributes to a better understanding of the role of LLMs like ChatGPT as potential assistants for portfolio managers.

KW - Asset selection

KW - ChatGPT

KW - Diversification effect

KW - FinTech

KW - Investment

KW - Large Language Model

KW - LLM

KW - Portfolio management

UR - https://www.scopus.com/pages/publications/85190996619

U2 - 10.1016/j.frl.2024.105433

DO - 10.1016/j.frl.2024.105433

M3 - Article

AN - SCOPUS:85190996619

SN - 1544-6123

VL - 64

JO - Finance Research Letters

JF - Finance Research Letters

M1 - 105433

ER -

Can ChatGPT improve investment decisions? From a portfolio management perspective

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this