Ooi Kok Loang, Sevenpri Candra

Machine Learning vs. Human Investors: Analyzing Adaptive Herding Behaviour in U.S. Stocks vs. Shariah-Compliant Stocks in Malaysia and Indonesia

Introduction

machine learning vs. Human investors: analyzing adaptive herding behaviour in u.s. Stocks vs. Shariah-compliant stocks in malaysia and indonesia. Analyze adaptive herding behavior in US & Shariah-compliant stocks (Malaysia/Indonesia) comparing machine learning vs. human investors. ML models outperform, offering insights for trading algorithms.

Abstract

This study examined the effectiveness of machine learning models in capturing adaptive herding behaviour in the US, Malaysia, and Indonesia. Utilising data from January 2010 to December 2023, the study incorporates market sentiment (Thomson Reuters MarketPsych Indices), news sentiment (Bloomberg Sentiment Analysis), and investor happiness measures (Hedonometer). The methodology employs both static and adaptive herding analyses using the CSAD approach, enhanced by real-time sentiment analysis and various machine learning models, including single- and multi-layer neural networks. The results indicate significant differences in herding behaviour across the three markets, with machine learning models demonstrating superior performance in capturing herding behaviour and faster normalisation after major macroeconomic events than traditional methods. These findings highlight the potential of machine learning models to challenge the static assumptions of the Efficient Market Hypothesis and provide insights for designing better trading algorithms by considering the impact of market sentiment, news sentiment, and investor happiness.

Review

This study presents a highly relevant and methodologically ambitious examination of adaptive herding behaviour, comparing the dynamics in US stocks with Shariah-compliant markets in Malaysia and Indonesia. By integrating a rich dataset spanning market, news, and investor happiness sentiment with advanced machine learning techniques, the authors aim to surpass traditional methods in capturing market inefficiencies. The abstract highlights a significant finding: machine learning models demonstrate superior capability in identifying herding and facilitating faster normalization after major events, thereby posing a direct challenge to the static assumptions underpinning the Efficient Market Hypothesis. A notable strength of this research lies in its innovative and comprehensive approach. The use of multiple real-time sentiment indicators—including Thomson Reuters MarketPsych, Bloomberg Sentiment Analysis, and the Hedonometer—provides a nuanced perspective on market psychology, moving beyond simple price-based analyses of herding. Furthermore, the application of both single and multi-layer neural networks within an adaptive CSAD framework represents a robust and cutting-edge methodology for analyzing complex market behaviors. The comparative analysis across a developed market and two distinct Shariah-compliant emerging markets is particularly valuable, offering unique insights into how cultural, regulatory, and ethical considerations might modulate herding dynamics, thereby contributing meaningfully to both behavioral finance and market efficiency literature. While the abstract provides a compelling overview, the full paper would benefit from elaborating on certain aspects. To fully substantiate the claims of "superior performance," the paper should detail the specific performance metrics used for the machine learning models and quantify their improvement over traditional methods. Similarly, a clearer definition of the "major macroeconomic events" analyzed and the precise methodology for measuring "faster normalisation" would enhance the clarity and replicability of the findings. Given the title's emphasis on "Human Investors," the paper could further explicitly link the machine learning models' insights to specific psychological biases or investor behaviors captured by the sentiment data, thereby providing more actionable guidance for the design of "better trading algorithms" and deepening the challenge to the Efficient Market Hypothesis.

Full Text

Comments

You need to be logged in to post a comment.