Porteføljeoptimering i pensionssektoren: En kvantitativ analyse af dynamisk allokering på tværs af aktivklasser i teori og praksis

Abstract

This thesis examines how pension funds can dynamically optimize portfolio allocations across equities, bonds, and alternative investments in a complex and uncertain economic environment. As long-term institutional investors, pension funds must continuously adjust their investment strategies in response to fluctuating interest rates, inflation pressures, and structural shifts in capital markets. The traditional reliance on government and corporate bonds has been challenged by prolonged periods of low interest rates, leading to a greater emphasis on alternative assets such as real estate and private equity. These asset classes offer diversification benefits and potential return enhancements, but they also introduce complications related to illiquidity, valuation uncertainty, and limited transparency.Using a combination of modern portfolio theory and the Black-Litterman model, the thesis investigates optimal asset allocations based on both historical returns and forward-looking return expectations. The analysis draws on monthly data from 1998 to 2024, covering twelve asset classes and capturing a wide range of market conditions including financial crises and monetary policy shifts. By applying rolling optimizations with a dynamic rebalancing approach, the study identifies how optimal asset weights evolve over time and how these differ from the actual allocation choices made by Danish pension funds. The findings highlight significant gaps between theoretical efficiency and institutional practice, often driven by real-world constraints such as regulation, transaction costs, and illiquidity considerations.The results suggest that while alternative investments can improve the risk return profile of a portfolio, their effective use requires a careful balance between theoretical allocation models and the operational realities of pension fund management. The inclusion of forward-looking market expectations enhances model robustness and provides a more realistic foundation for long-term allocation strategies. Ultimately, the thesis demonstrates the value of integrating quantitative models with institutional insight to support resilient and adaptive investment decisions in an increasingly volatile financial landscape.