Challenging the Default : A Stochastic Approach of Pension Portfolio Glide Paths

Sammanfattning

The thesis evaluates both rising and declining equity glide path strategies for Swedish pensioners during the decumulation phase compared to AP7 Såfa. The AP7 Såfa enters the decumulation phase at approximately 57\% equity and declines linearly to a fixed floor of 35\%. A stochastic simulation framework is constructed using historical Swedish stock and bond real return data. It combines AR and VAR conditional mean models, GARCH volatility models, and a pairwise filtered historical simulation bootstrap. Four simulation model specifications of increasing complexity are estimated and used, each generates 10,000 different return paths of a 30-year horizon. Three rising and one decreasing equity glide path are evaluated against AP7 Såfa using success rate, median terminal wealth and median depletion age at a 4\% withdrawal rate. Under average market conditions, all strategies achieve high success rate, as the sole meaningful difference emerges under unfavourable sequence of return risk periods. When markets perform poorly in the early stages of the decumulation phase, AP7 Såfa consistently achieves the lowest success rate, ranging from 92.0\% to 92.9\%, compared to 94.5\% to 96.3\% for the custom equity profiles. Among the evaluated profiles, Profile 3 with a terminal equity allocation of 60\% achieves the most favourable balance between downside protection and long-term portfolio growth. These findings are robust across all four model specifications, indicating that the conclusions reflect characteristics of the Swedish return data rather than a single modelling assumption.