Regime-Switching Signals for Pairs Trading Through Structural Breaks

Summary

This document describes a pairs trading method that uses a two-state Markov regime-switching model to assess whether spread deviations may reflect a persistent change rather than temporary mean reversion. The proposed signal combines the estimated regime and its smoothed probability with a deviation threshold. In a high-mean regime, a long spread position is considered when the spread falls sufficiently below its mean and the probability of remaining in that regime clears a chosen threshold; the low-regime logic reverses this direction.

The workflow selects candidate pairs, constructs a spread, fits the model on a rolling window, derives a signal, and combines it with the existing position to decide whether to trade. The article motivates the method through prior research and provides implementation steps, but reports no backtest results. Its limits include the possibility of future regimes beyond the two observed states, which could defeat break detection, and fitting difficulties when both the mean and variance are allowed to switch. A stricter signal rule is suggested to reduce exposure to unseen extreme regimes.

Key ideas

• A Markov regime model can help distinguish temporary spread moves from structural changes.
• Signals combine the spread deviation with the estimated regime and its smoothed probability.
• The outlined process covers pair selection, spread construction, rolling estimation, and trade decisions.
• Unseen regimes beyond the model’s observed states can still cause large losses.
• Allowing both the process mean and variance to switch can make model fitting fail.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.