PFP2433. A Random Walk Towards More Efficient (Retirement) Portfolios

Investment returns are commonly assumed to be independent (or random) across time in financial models. In reality, investments have historically exhibited varying levels of serial dependence, where the returns evolve nonrandomly for both individual investments (i.e., autocorrelation) as well as across investments. This presentation explores how these effects impact optimal portfolio allocations and notes considerable differences across investment horizons, which can be especially pronounced for more risk averse investors who are concerned with inflation risk (e.g., retirees). Overall, this research suggests investment professionals need to more actively consider serial dependence when building portfolios to ensure they are best aligned to help clients accomplish their goals.

Learning Objectives:

• Assess how risks can potentially change over time.
• Identify more efficient portfolios for clients for different time horizons based on the unique relationships that have existed among assets and risk across time.