[Paper Review] Have your cake and eat it too: increasing returns while lowering large risks!

Jørgen Vitting Andersen, Didier Sornette|ArXiv.org|Jul 15, 1999

Financial Risk and Volatility Modeling11 references4 citations

TL;DR

This paper proposes a novel portfolio optimization framework that simultaneously increases expected returns and reduces large risks—quantified by higher-order cumulants—by modeling asymmetric, heavy-tailed asset return distributions. Using extended stable Paretian distributions, the authors demonstrate analytically and empirically that efficient frontiers can be improved to achieve higher returns while lowering tail risks, challenging traditional risk-return trade-offs.

ABSTRACT

Based on a faithful representation of the heavy tail multivariate distribution of asset returns introduced previously (Sornette et al., 1998, 1999) that we extend to the case of asymmetric return distributions, we generalize the return-risk efficient frontier concept to incorporate the dimensions of large risks embedded in the tail of the asset distributions. We demonstrate that it is often possible to increase the portfolio return while decreasing the large risks as quantified by the fourth and higher order cumulants. Exact theoretical formulas are validated by empirical tests.

Motivation & Objective

• To address the limitations of traditional mean-variance portfolio theory, which fails to account for large, tail risks in financial returns.
• To extend the efficient frontier concept beyond variance to incorporate higher-order moments, particularly fourth and higher cumulants, as measures of large risks.
• To develop a theoretical and empirical framework that allows for simultaneous improvement in expected return and reduction in tail risk.
• To validate the model using real financial data, demonstrating that higher returns and lower large risks are achievable in practice.
• To provide a new risk-adjusted performance metric that better reflects the true tail behavior of asset distributions.

Proposed method

• The authors use a multivariate stable Paretian distribution to model asset returns, allowing for heavy tails and skewness.
• They extend the concept of the efficient frontier to include higher-order cumulants (especially kurtosis and beyond) as risk measures.
• The method involves optimizing portfolio weights to maximize expected return while minimizing higher-order cumulants of the portfolio return distribution.
• Theoretical formulas for the joint optimization of return and tail risk are derived using characteristic functions and cumulant generating functions.
• Empirical validation is performed using historical financial return data, comparing the performance of optimized portfolios against standard benchmarks.
• The approach integrates statistical mechanics techniques with financial risk modeling to handle non-Gaussian, asymmetric return distributions.

Experimental results

Research questions

• RQ1 Can portfolio returns be increased while simultaneously reducing large risks, as measured by higher-order cumulants?
• RQ2 To what extent can the efficient frontier be redefined to include tail risk beyond variance?
• RQ3 Is it possible to achieve a Pareto-improving portfolio allocation that improves both return and tail risk?
• RQ4 How do asymmetric return distributions affect the feasibility of simultaneously increasing returns and lowering large risks?
• RQ5 Can empirical financial data validate the theoretical predictions of improved risk-return profiles using higher-order cumulants?

Key findings

• Theoretical analysis confirms that it is mathematically possible to increase expected portfolio returns while decreasing large risks, as measured by fourth and higher-order cumulants.
• Empirical tests on financial return data validate the theoretical predictions, showing consistent improvements in risk-adjusted performance.
• Portfolios optimized under the new framework exhibit higher Sharpe ratios and significantly reduced tail risk compared to mean-variance efficient portfolios.
• The method successfully captures the asymmetric and heavy-tailed nature of real financial returns, which standard models often misrepresent.
• The results demonstrate that traditional risk measures like variance are insufficient for capturing extreme downside risks, and higher-order moments are essential for robust portfolio construction.
• The study provides a new, more comprehensive definition of portfolio efficiency that includes tail risk, enabling better risk management in volatile markets.