ACD Modeling High-Frequency FX and Market Microstructure

Abstract

This paper advances high-frequency foreign exchange (FX) market microstructure analysis by adapting Autoregressive Conditional Duration (ACD) models to study intervals between price updates. By treating these updates as random variables within a point process, the models adeptly capture the dynamic structure of conditional durations and retain key information in high-frequency series. These series display properties critical for understanding market behavior and liquidity dynamics. The findings challenge the belief that increased data frequency reduces microstructural relevance, showing it actually improves understanding of market dynamics. This study broadens econometric model applications and offers updated insights into FX market behavior, providing practical information for academics, practitioners, and policymakers. It contributes significantly to the literature and lays a foundation for future research.

Keywords

• High-Frequency
• Market Microstructure
• Stylized Facts
• Point Processes
• Autoregressive Conditional Duration Models
• Price Durations