When the Rules Change: Adaptive Signal Extraction via Kalman Filtering and Markov-Switching Regimes

TL;DR

This study tackles the instability of order-flow–returns relationships under market stress by developing a dynamic, state-dependent framework that adapts to shifting conditions. It combines an Adaptive Kalman Filter with volatility-driven measurement noise, a three-state Markov-Switching regime model, and an asymmetric response function to extract signals across Bull, Normal, and Crisis periods in the Korean stock market. The main contributions are methodological—showing how to integrate state-space techniques to accommodate regime changes—and empirical—revealing that foreign investor predictive power surges during crises ($β_{crisis} = 0.00204$) while retail traders exhibit momentum-driven responses. The integrated All-Weather strategy yields modest drawdown relief during extreme events, with notable performance in 2020 but mixed results in later periods, underscoring the importance of regime-aware risk management for practitioners and regulators.

Abstract

Static linear models of order flow assume constant parameters, failing precisely when they are needed most: during periods of market stress and structural change. This paper proposes a dynamic, state-dependent framework for order flow signal extraction that adapts to shifting market conditions in the Korean stock market. Using daily transaction data from 2020--2024 covering 2,439 stocks and 2.79 million stock-day observations, we implement three complementary methodologies: (1) an Adaptive Kalman Filter where measurement noise variance is explicitly coupled to market volatility; (2) a three-state Markov-Switching model identifying Bull, Normal, and Crisis regimes; and (3) an Asymmetric Response Function capturing differential investor reactions to positive versus negative shocks. We find that foreign investor predictive power increases 8.9-fold during crisis periods relative to bull markets ($β_{crisis}=0.00204$ vs. $β_{bull}=0.00023$), while individual investors exhibit momentum-chasing behavior with 6.3 times stronger response to positive shocks. The integrated ``All-Weather'' strategy provides modest drawdown reduction during extreme market events, though challenges remain in the post-COVID high-rate environment.

Figures (10)

• Figure 1: Kalman Gain versus Market Volatility
• Figure 2: Time Series of Regime Probabilities
• Figure 3: Asymmetric Response Patterns by Investor Type
• Figure 4: Cumulative Returns: Raw vs. Filtered vs. All-Weather
• Figure 5: Comparison of Filtered versus Raw Order Flow Signal