Multi-Regime, Goal-Based Retirement Solutions: Sensitivity Analysis and Post-Retirement Performance Comparison of Dynamic Strategies

Abstract

The shift from defined-benefit to defined-contribution pension plans worldwide has left many individuals who lack access to private banking services without proper retirement financial planning. As individuals have more responsibilities for their retirement planning, there is a growing concern of choosing asset allocation strategies that are not suitable for their goals.

Besides, in the light of the current situation regarding the COVID-19 pandemic, individuals must be able to prepare for a possible prolonged “crash” economic regime. Understanding the impact of such a crash regime is important for individuals to make optimal financial decisions. These aspects will be considered in the paper.

The objective of this study is to model and comprehend a fully automated retirement planning system for individual investors. By understanding the characteristics of each strategy through various risk measures and analyzing its sensitivity to prolonged crash economic regimes, health risks, savings rate, and replacement income ratio (spending after retirement, divided by the last salary before retirement), we can use the results to recommend the most appropriate asset allocation and savings rate strategy for an individual investor.

In this study, we implement multi-regime Monte Carlo simulations to simulate asset returns throughout the accumulation phase (pre-retirement) and decumulation phase (post-retirement). We consider different asset allocation strategies: standard glide path (SGP: linearly reduce equity weight over time), reversed glide path (RGP: linearly increase equity weight over time), “Catch-Up” strategy (CGP: follow SGP and increase equity weight when the wealth falls behind a benchmark portfolio, called “goal portfolio”), “Play-Safe” strategy (PGP: follow SGP and decrease equity weight when the wealth exceeds the goal portfolio), and “Dynamic Zone” glide path (DGP: combine CGP and PGP). We analyze the performances concerning non-goal-based and goal-based risk measures while incorporating the longevity risk and its corresponding risk measures, such as the probability of outliving retirement savings, to further comprehend the characteristics of each policy. We evaluate a goal-based, dynamic strategy that alters the savings rate based on its performance relative to the goal portfolio. That is, it increases the savings rate during a crash regime and decreases the savings rate during a normal regime. Moreover, we conduct sensitivity analyses on 4 parameters: prolonged crash economic regime, health risks, savings rate, and replacement income ratio.

The results are analyzed to fully comprehend the strengths and weaknesses of each strategy. First, we recommend PGP for individual investors, especially when there is a prolonged crash regime. Generally, it outperforms other strategies in average maximum drawdown ratio, Goal-at-Risk (GaR), Conditional Goal-at-Risk (CGaR), and the probability of outliving retirement savings. If a 10-year crash regime starts after the middle of the pre-retirement phase, PGP best protects the portfolio from a huge loss during the crisis as it has the lowest probability of outliving retirement savings, which is 18 percentage points lower than that of SGP. With PGP, individuals need to increase their savings rate by 8-31 percentage points, depending on the timing of a 10-year crash regime, to narrow the probability of outliving savings to 10%. However, it is less effective to increase the savings rate only during the crisis or to decrease the replacement income ratio.

Besides, for any allocation strategy, the worst time to have a prolonged crash regime is in the middle of the pre-retirement phase, leading to an increase of 69 percentage points in the probability of outliving savings. In terms of savings rate, shifting more savings to an earlier part in the pre-retirement phase can result in a reduction in both the average maximum drawdown and the probability of outliving retirement savings for CGP, PGP, and DGP. Also, the result of the goal-based, dynamic savings rate strategy, in terms of the probability of outliving savings, is equivalent to that of increasing the savings rate for the entire pre-retirement period by 12%.