Measuring Sequence of Returns Risk

Abstract

We discuss the nature and importance of the concept of Sequence Risk, the risk that a bad return occurs at a particularly unfortunate time, such as around the point of maximum accumulation or the start of decumulation. This is especially relevant in the context of retirement savings, where the implications for withdrawal rates of a bad return can be particularly severe. We show how the popular 'glidepath' or target date savings' products are very exposed to such risk. Three different measures of Sequence Risk are proposed, each of which is intended to inform investors of the probability that a chosen investment strategy may not deliver desired withdrawal rates and hence these measures are intended to aid investment choices; conventional performance measures such as Sharpe or Sortino ratios are only indirectly related to this ability to achieve a given withdrawal experience. Finally, we note that, using US data, very simple portfolios comprising equities and bonds can achieve very low probabilities of failure to achieve popular desired withdrawal rates such as 5% p.a. providing the equity component is 'smoothed' by switching in and out of cash using a simple trend following rule.

Optimal withdrawal frequency for sustainable retirement withdrawals

“…Noting the importance of return volatility, authors have explored the influence of various asset allocations (e.g., Clare, Glover, et al, 2021; Estrada, 2018) and investment strategies (e.g., Clare et al, 2017; Milevsky & Posner, 2014) on retirement withdrawals sustainability. Spitzer et al (2007) summarize much of the evidence regarding new retirees with a 30‐year planning horizon and conclude that a new retiree who has at least 50% stock exposure can withdraw 4% to 4.5% of the portfolio in the initial year and an inflation‐adjusted equivalent amount each year thereafter with 90% to 95% confidence that the portfolio will not run out of money within 30 years.…”

“…Many authors have also highlighted the importance of the sequence of return risk (i.e., large negative returns in the beginning years of a withdrawal program, such as occurred in 1973–1974) to the sustainability of a retirement withdrawal program (e.g., Clare, Glover, Seaton, Smith, and Thomas 2021; Estrada, 2018; Finke et al, 2013; Kenigsberg et al, 2014; Milevsky & Posner, 2014). The risk of a retirement withdrawal program failing (i.e., running out of money before the end of the investment horizon) is directly related to beginning retirement withdrawals prior to a bear market.…”

“…1 1 Examples of authors modeling retirement distributions at annual intervals include Bengen (1994, 1996, 1997, 2001), Cooley et al (1998, 1999, 2001, 2003b), Pye (2000, 2001), Milevsky and Robinson (2005), Milevsky and Posner (2014), Milevsky and Huang (2011), Waring and Siegel (2015), Pye (2000, 2001), Suarez et al (2015) Suarez (2020), Reichenstein (2006), Guyton and Klinger (2006), Blanchett et al (2012), Blanchett (2023), Pfau (2011), Stout (2008), Estrada (2018, 2021), Laster, Suri and Vrdoljak (2012), Kenigsberg et al (2014), Clare et al (2017), Clare, Glover Seaton, Smith and Thomas (2021a), Xu (2018) Sun and Lan (2022). …”