Dynamic Asset Allocation with Expected Shortfall via Quantum Annealing.

Area of Science:

• Quantum Computing
• Financial Optimization
• Computational Finance

Background:

• Classical optimization methods struggle with computationally expensive financial problems.
• Quantum hardware advancements offer novel solutions for complex optimization tasks.
• Modeling extreme market events requires sophisticated risk metrics like expected shortfall.

Purpose of the Study:

• To propose a hybrid quantum-classical algorithm for dynamic asset allocation.
• To incorporate a target return and expected shortfall risk metric into portfolio optimization.
• To formulate the Markowitz optimization as a Quadratic Unconstrained Binary Optimization (QUBO) problem for quantum processing.

Main Methods:

• An iterative algorithm was developed, treating target return as a constraint within a Markowitz portfolio optimization framework.
• The algorithm dynamically adjusts target return to meet specified expected shortfall targets.
• Markowitz optimization was reformulated as a QUBO problem, suitable for quantum annealers.

Main Results:

• The hybrid algorithm was tested on D-Wave's 2000Q and Advantage quantum annealers using real-world financial data.
• Quantum annealers generated portfolios achieving over 80% of classical optimal returns while meeting expected shortfall targets.
• Higher asset correlations in experiments correlated with improved performance, suggesting practical near-term quantum applications.

Conclusions:

• Hybrid quantum-classical approaches show promise for solving complex financial optimization problems.
• Quantum annealers can effectively manage portfolio risk, even during extreme market events.
• Further research into asset correlation may enhance the design of practical quantum financial applications.