Exploring Quantum-Enhanced Estimation of Financial Risk Metrics with Quantum RNG

TL;DR

The paper tackles accurate estimation of $VaR_{\alpha}$ and $CVaR_{\alpha}$ in risk management by leveraging quantum-generated randomness to enhance Monte Carlo simulations. It presents a Quantum Random Numbers Generation (QRNG) framework that can be realized via photonic RNG or quantum processing units to produce true random samples for Monte Carlo, and it assesses the resulting impact on estimation accuracy. Empirical results on a 40-asset portfolio show that quantum-based randomness yields slightly lower risk estimates and reduced estimator variance compared with classical pseudo-random sampling, demonstrated on a 2-day horizon with $2\times 10^6$ paths. Benchmarking confirms higher entropy and reasonable independence across RNG sources, supporting practical viability on a risk platform (Scenario X) for scalable, quantum-assisted risk assessment.

Abstract

In this paper, we present an approach for estimating significant financial metrics within risk management by utilizing quantum phenomena for random number generation. We explore Quantum-Enhanced Monte Carlo, a method that combines traditional and quantum techniques for enhanced precision through Quantum Random Numbers Generation (QRNG). The proposed methods can be based on the use of photonic phenomena or quantum processing units to generate random numbers. The results are promising, hinting at improved accuracy with the proposed methods and slightly lower estimates (both for VaR and CVaR estimation) using the quantum-based methodology.

Figures (6)

• Figure 1: The flow chart of a Quantum Random Number Generator that exploits a single QPU.
• Figure 2: The architecture for quantum random number generation with the quantum circuit.
• Figure 3: Histograms deriving from the uniformity tests for the three methodologies implemented
• Figure 4: Scenario X SaaS Platform
• Figure 5: Returns simulated using the random numbers provided by ANU.