An Entropic Approach for Pair Trading in PSX

Area of Science:

• Quantitative Finance
• Financial Risk Management
• Algorithmic Trading

Background:

• Pair trading relies on the principle of price convergence between cointegrated assets.
• Identifying optimal trade entry/exit points and managing model misspecification risks are significant challenges in pair trading.
• Existing approaches may not adequately address uncertainty and model errors, potentially leading to financial losses.

Purpose of the Study:

• To introduce a novel entropic approach for pair trading.
• To utilize entropy as a penalty function to mitigate model misspecification risks.
• To evaluate the performance of the entropic approach against established trading strategies.

Main Methods:

• Employed a novel entropic approach incorporating entropy as a penalty function for model misspecification.
• Utilized daily stock data for 64 companies listed on the Pakistan Stock Exchange (PSX) from 2017-2019.
• Computed stock returns using the entropic approach and compared them against a buy and hold strategy.

Main Results:

• The entropic approach demonstrated positive and statistically significant returns in pair trading.
• Pair trading using the entropic method showed a performance advantage over the buy and hold strategy.
• The entropic approach outperformed distance-based and machine learning approaches in the Pakistani market context.

Conclusions:

• The entropic approach offers a promising new methodology for pair trading.
• Entropy serves as an effective risk measure and penalty function in financial modeling.
• The entropic approach presents a competitive alternative to existing methods for the Pakistani stock market.