Market efficiency and long-term memory: analyzing financial markets through the Hurst exponent

• Ano de defesa: 2024
• Autor(a) principal: Gasparotto, Angelica Castilho
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Análise de Alcance Reescalonado (R/S); Crises financeiras; Eficiência do mercado; Expoente de Hurst; Financial crises; Hurst exponent; Long memory; Market efficiency; Memória longa; Rescaled Range (R/S) analysis; Retornos; Returns; Volatilidade; Volatility
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/96/96133/tde-20022025-090058/

This dissertation investigates the application of the Hurst exponent as a tool for understanding market efficiency, persistence, and structural changes in financial markets, particularly during periods of economic turbulence. The Hurst exponent, which quantifies long-term memory in time series, serves as an alternative to the Efficient Market Hypothesis (EMH), enabling the assessment of market behaviors beyond random walks. The study explores two core approaches: first, examining the Hurst exponent across major financial indices (Ibovespa, FTSE, S&P 500, and Dow Jones) during significant crises, including the Dot-Com Bubble, the Financial Crisis of 2008, and the COVID-19 crisis; and second, analyzing the Hurst exponent over 21-day intervals to correlate with short-term volatility and returns. Employing Rescaled Range (R/S) analysis as the primary methodological tool, this research demonstrates the ability of the Hurst exponent to capture persistent behaviors and trend-following tendencies within financial time series, which deviate from traditional EMH assumptions. The findings reveal that, during crisis periods, financial indices exhibit high Hurst values, indicating persistent market behavior and potential predictability amidst instability. Additionally, the analysis shows that the Hurst exponent correlates moderately with returns but less so with volatility over short-term intervals, suggesting its capacity as a trend indicator rather than a measure of immediate risk. This study contributes to the literature by highlighting the practical utility of the Hurst exponent in both crisis prediction and market analysis, supporting its role as a complementary tool alongside traditional financial metrics. Future research could further explore the predictive applications of the Hurst exponent across diverse asset classes and under varying economic conditions.