Inverse Optimization: A New Perspective on the Black-Litterman Model.

Area of Science:

• Quantitative Finance
• Financial Modeling
• Asset Allocation

Background:

• The Black-Litterman (BL) model is a standard in financial asset allocation.
• Its traditional framework relies on statistical methods and mean-variance optimization.

Purpose of the Study:

• To present a novel perspective on the Black-Litterman model using inverse optimization.
• To expand the model's applicability by incorporating investor information on volatility and market dynamics.
• To extend the framework beyond mean-variance optimization to include coherent risk measures.

Main Methods:

• Replacing the statistical framework of the original BL model with inverse optimization principles.
• Developing new "BL"-type estimators: Mean Variance Inverse Optimization (MV-IO) and Robust Mean Variance Inverse Optimization (RMV-IO) portfolios.
• Utilizing ideas from arbitrage pricing theory and volatility uncertainty.

Main Results:

• The proposed inverse optimization approach offers a richer formulation for the BL model.
• New estimators (MV-IO, RMV-IO) are computationally introduced and studied.
• Numerical simulations and historical backtesting demonstrate improved risk-reward tradeoffs compared to traditional BL methods.
• The new approaches show enhanced robustness against inaccurate investor views.

Conclusions:

• Inverse optimization provides a powerful alternative framework for the Black-Litterman model.
• The developed MV-IO and RMV-IO portfolios offer superior performance and robustness.
• This research broadens the scope of asset allocation models to include advanced risk measures and market dynamics.