Dynamic vine copula models for high-frequency data

Dynamic vine copula models for high-frequency data

• Author
• Murilo Andre Peres Pereira
• Abstract

• This study investigates intraday patterns in the comovements of financial stock returns, focusing on the importance of flexible dependence structures on density forecasting. We propose a dynamic canonical vine copula method, which models complex dependence patterns, including both time-varying and asymmetric dependencies in the upper and lower tails among financial assets. Utilizing a pair copula decomposition approach, this research analyzes 1-minute frequency returns of 10 U.S. financial stocks in March 2020, a period marked by Covid-19 market turmoil. Our findings highlight the critical role of tail dependencies and time-varying parameters in accurately modeling and forecasting intraday returns.

• Keywords
• : forecasting; regular vine; pair-copula constructions; time-varying copulas;
• Modality
• Comunicação oral
• Subject Area
• Engenharia Financeira (Financial Engineering)

Back Download