T. Rowe Price: Beyond sequencing risk: Dynamic withdrawals for retirement

Key Insights

• Sequencing risk—the threat of market downturns early in retirement—can damage retirement income by reducing future portfolio growth potential.
• One key to managing sequencing risk is implementing an income strategy that seeks to limit portfolio withdrawals when market returns are poor.
• Our analysis of historical market performance since 1991 suggests a dynamic strategy that varies withdrawals based on market conditions can improve outcomes.

One of the most pertinent investment risks that retirees face is sequencing risk—the danger that poor market conditions early in retirement will deplete their wealth at a faster‑than‑expected rate as low or negative returns coincide with portfolio withdrawals.

The negative impact of sequencing risk—also known as sequence‑of‑returns risk—on retirement outcomes can extend far beyond the shorter‑term volatility associated with events such as the COVID‑19 pandemic or the more recent tariff‑related market sell-off.

By forcing recent retirees to sell more assets to meet their immediate income needs, poor returns caused by market downturns may reduce the capital available for future portfolio growth during potential market rebounds. The result can be a permanent reduction in the portfolio’s ability to generate retirement income. In contrast, market downturns later in retirement typically are less damaging, as the portfolio already has supported withdrawals over time.

We believe that managing sequencing risk effectively requires a nuanced approach that balances income needs with capital preservation. This involves more than just careful portfolio construction and ongoing asset allocation management; it also requires a thoughtful withdrawal strategy that considers the investor’s financial circumstances and tolerance for risk.

Sequencing risk can significantly impact outcomes

Figure 1 illustrates the significant effect that sequencing risk can have on retirement outcomes. The charts compare the patterns of annual returns experienced by two hypothetical retirees and how those patterns could affect their portfolio balances over time.

• Each hypothetical investor starts with a GBP 100,000 portfolio and withdraws GBP 7,000 annually.
• While the hypothetical annual returns are equivalent for both investors, the sequence of those returns differs dramatically in the first 10 years.
• As a result, while both hypothetical investors experience portfolio declines, the effect on their respective portfolio balances is very different.
• The hypothetical investor who faced a significant market downturn early in retirement depleted their savings much faster than the investor who encountered a comparable decline later.

Annual returns matter, but the sequence can matter more

(Fig. 1) Hypothetical portfolio returns and balances over a 15year retirement window.

Portfolio Balances

Figures refer to simulated past performance and that past performance is not a reliable indicator of future performance.
For illustrative purposes only. Please see the Appendix for additional important information on this analysis.
¹ Annual returns shown are purely hypothetical and not based on the historical market performance of any asset class. All figures calculated in GBP.
Data analysis by T. Rowe Price.

We believe that by understanding and planning for sequencing risk, investors can better protect their retirement savings from the potentially devastating effects of poor market returns early in retirement. Investors and/or their financial advisors should review this approach regularly.

Common strategies for managing sequencing risk

Three strategies are widely used in the UK to mitigate sequencing risk. These are generally known as the cash buffer, bucketing, and dynamic withdrawal approaches.

• Cash buffer. By maintaining a significant cash reserve (typically equal to two to three years of expenses) to weather market downturns, retirees may be able to avoid selling assets when prices are temporarily depressed. However, one drawback of this approach is that large cash reserves may lead to slower portfolio growth, as long‑term returns for cash may not be as high as for other investments.
• Bucketing (also known as horizon planning). By allocating funds into low‑risk investments for short‑term needs and higher‑risk investments for longer‑term needs, investors can adjust the sources of their withdrawals based on market conditions. A potential drawback is the complexity of managing multiple investment buckets, which may require frequent monitoring and adjustments.
• Dynamic withdrawal. In this approach, investors make larger withdrawals from their portfolios when market conditions are good and smaller ones in bad years. A key benefit of this approach is that the balance trajectory can be managed to keep it from drifting too high or too low because of the withdrawal adjustments. The downside of this approach is that income fluctuates and becomes less predictable.

Under both the cash buffer and bucketing approaches, retirees withdraw a fixed amount each year to cover living expenses. This can provide stability and predictability and offer greater confidence in meeting income needs. By contrast, the stability and predictability of income under the dynamic withdrawal approach will depend heavily on how the withdrawal strategy is designed. This creates complexity and typically requires careful research and expert guidance to succeed.

One simplified version of dynamic withdrawal is the “4% rule” proposed by U.S. financial planner William Bengen as a rule of thumb for determining how much a retiree should withdraw from a retirement account each year. Retirees using this rule set their withdrawals to equal 4% of the value of their portfolios and then maintain that same percentage—rather than a fixed amount—going forward.1 Some financial advisers in the UK have adopted this guideline.

Other methods exist to mitigate sequencing risk, such as phased retirement, in which investors gradually transition into retirement by reducing work to manage the income pressure on their portfolios. However, phased retirement is, in essence, a personal choice and can be combined with the strategies outlined above.

Our analysis favours a dynamic withdrawal approach

To explore the potential for dynamic withdrawal strategies, we developed a generalised dynamic withdrawal approach and then analysed the potential impact on hypothetical retirement portfolios.

Our generalised approach considered multiple parameters to guide the withdrawal process, including long‑term expected returns for the hypothetical portfolios, historical market returns, and a preferred set of income and balance trajectory profiles. In our analysis, this approach resulted in larger income payouts when recent market returns were higher than long‑term expected returns, or if a higher income profile was preferred.

We assumed that our hypothetical investor retired with a GBP 500,000 accumulated portfolio balance and sought to generate an annual income of about GBP 31,3002 per year. We then compared investor experiences over 15‑year retirement periods under the cash buffer and bucketing approaches, as well as under our own dynamic withdrawal approach.

• Cash buffer: We assumed our hypothetical retiree allocated two years of expenses to cash and invested the remaining balance in a balanced portfolio of 50% stocks and 50% bonds.3 If the previous year’s price return for the balanced portfolio was negative, we assumed the retiree withdrew from their cash reserve instead until it was depleted.
• Bucketing: This hypothetical investor set aside a cash reserve equivalent to two years of expenses and allocated three years of spending to a lower‑risk portfolio invested 20% in equities and 80% in bonds, while placing the remaining balance in a higher‑risk portfolio invested 70% in stocks and 30% in bonds. If the previous year’s return for the high‑risk portfolio was negative, we assumed the investor withdrew first from their cash reserve, then from the lower‑risk portfolio, and finally from the higher‑risk portfolio. We also assumed that withdrawals were made sequentially—with each bucket fully depleted before moving to the next.
• T. Rowe Price’s dynamic withdrawal approach: All savings were invested in a balanced portfolio allocated 50% to stocks and 50% to bonds. The investor started with a 6.26% withdrawal rate in order to equal the initial withdrawal amount of GBP 31,300 based on the assumed GBP 500,000 starting balance. Subsequent yearly withdrawals were adjusted dynamically to reflect portfolio annual returns, long‑term expectations, and the desired income and portfolio balance profiles.

Investors were assumed to retire at the start of each calendar year from 1991 (the earliest available data for the bond index) to 2010, creating, in effect, a series of rolling 15‑year performance windows.

To assess how well each approach potentially mitigated sequencing risk, we first configured our dynamic withdrawal strategy to closely match the fixed withdrawal patterns used in the cash buffer and bucketing approaches. While fixed withdrawals increase the risk of portfolio depletion in extended market downturns, the past three decades have seen few such prolonged declines. None of the hypothetical portfolios were fully depleted over our assumed investment time periods.

We calculated hypothetical portfolio balance and income trajectories over the 15‑year retirement windows using the three approaches outlined above, based on historical annual stock, bond, and cash returns. Our analysis compared portfolio balances at the end of each 15‑year window, with income levels aligned. A higher ending balance indicated better resilience against sequencing risk.

Dynamic withdrawals delivered higher balances

In our analysis, the dynamic withdrawal approach consistently achieved higher ending balances compared with both the cash buffer and the bucketing methods, regardless of the retirement window examined (Figure 2). On average over all the rolling 15‑year windows, the dynamic approach delivered 7% higher balances than the cash buffer and 13% higher than the bucketing approach.

A dynamic withdrawal approach can improve outcomes

(Fig. 2) Hypothetical portfolio balances under three withdrawal approaches.

Figures refer to simulated past performance and that past performance is not a reliable indicator of future performance.
For illustrative purposes only. Please see the Appendix for the return assumptions and for additional important information on this analysis.
Starting balance = GBP 100,000 and starting annual withdrawal = GBP 31,300. All figures calculated in GBP.
Source: Bloomberg Finance L.P. Data analysis by T. Rowe Price.

In the 15‑year retirement windows where hypothetical end balances were lowest—in other words, where sequencing risk was most pronounced—our dynamic withdrawal approach produced the largest differences.

For example, for individuals who retired in 2000 and 2001 (and thus were hit both by the collapse of the dot-com bubble and the 9/11 market correction), our approach resulted in approximately 15% higher balances compared with the cash buffer method and just over 40% higher relative to the bucketing approach.

Our enhanced dynamic approach improved income

While sequencing risk is an important component of retirees’ retirement experience, it primarily focuses on the risk of balance depletion. T. Rowe Price takes a more holistic approach to enhance retirees’ overall experience. We have developed a five‑dimensional (5D) framework to analyse retiree preference and create solutions.4 In addition to sequencing risk, our framework considers the income side of the puzzle, examining both income levels and income volatility.

To further demonstrate the benefits of our dynamic withdrawal approach, we extended the analysis to include a variation on our initial methodology, one that focuses on both capital preservation and income profiles. We call this the “TRP dynamic enhanced” approach.

In this configuration, we allowed the income stream to exhibit greater volatility. However, the increased volatility seen in our hypothetical portfolios primarily occurred in strong market periods. Conversely, in periods with weak market returns—when balances were lower and the risk of balance depletion higher—increases in income volatility were relatively minimal (Figure 3).

Withdrawal approaches can be adjusted for different risk tolerances

(Fig. 3) Hypothetical income volatilities based on withdrawal approach.

Figures refer to simulated past performance and that past performance is not a reliable indicator of future performance.
For illustrative purposes only. Please see the Appendix for the return assumptions and for additional important information on this analysis.
All figures calculated in GBP. Volatility was defined as the standard deviation of annual income, normalised by average income over each of the
15‑year windows.
Source: Bloomberg Finance L.P. Data analysis by T. Rowe Price. 

It is also worth noting that despite the increased income fluctuations, the volatility of income streams across most retirement windows remained within the 10% tolerance range shown in Figure 3.

The main advantage of allowing slightly higher income volatility was that our hypothetical retirees achieved higher annual incomes and thus could spend more if they so desired. Over most of the 15‑year windows shown in Figure 4, retirees increased their average incomes, with those gains ranging from just over GBP 100 per year to as much as GBP 7,000 per year—with an average improvement of 8%. In the one window in which retirees received a lower average annual income, the decrease was minimal, at just under GBP 250.

Retirees willing to accept higher volatility may be able to boost income

(Fig. 4) Hypothetical average annual incomes across 15year retirement periods.

Figures refer to simulated past performance and that past performance is not a reliable indicator of future performance.
For illustrative purposes only. Please see the Appendix for the return assumptions and for additional important information on this analysis.
Starting balance = GBP 100,000 and starting annual withdrawal = GBP 31,300. All figures calculated in GBP.
Source: Bloomberg Finance L.P. Data analysis by T. Rowe Price.

In our analysis, the income increases made possible by our enhanced dynamic withdrawal methodology did not compromise the approach’s strong ability to mitigate sequencing risk. The higher income arose primarily in periods when market returns were favourable and more robust portfolio balances could sustain greater income levels.

In windows with more adverse market returns, where hypothetical portfolio balances were on the lower end, average incomes remained largely unchanged or were slightly reduced in our analysis.

In the two most challenging retirement windows shown in Figure 5—the ones beginning in 2000 and 2001—the enhanced approach resulted in higher balances, indicating a reduced risk of portfolio depletion and more effective mitigation against sequencing risk.

An enhanced dynamic approach improved balances even in extreme cases

(Fig. 5) Hypothetical ending portfolio balances across 15year retirement periods.

Figures refer to simulated past performance and that past performance is not a reliable indicator of future performance.
For illustrative purposes only. Please see the Appendix for the return assumptions and for additional important information on this analysis.
Starting balance = GBP 100,000 and starting annual withdrawal = GBP 31,300. All figures calculated in GBP.
Source: Bloomberg Finance L.P. Data analysis by T. Rowe Price.  

Conclusions

We believe that investors seeking secure and fulfilling retirement outcomes need to consider sequencing risk as part of their investment planning. In the analysis above, we examined some of the challenges of addressing sequencing risk, highlighting how market volatility can significantly impact retirees’ financial well‑being.

In our analysis, the dynamic withdrawal approach adapted to changing market conditions, delivered higher average incomes, and maintained more robust portfolio health—unlike traditional methods, such as cash buffer and bucketing, using either constant pound or constant rate withdrawals. The objectives of the dynamic approach are to balance income streams and safeguard and grow portfolio balances, providing a smoother financial experience for our hypothetical retirees.

We believe that embracing a dynamic withdrawal strategy potentially can allow retirees to manage their savings more confidently and improve their chances of enjoying the lifestyles they envision, possibly alleviating some of the uncertainties of retirement and supporting a course toward financial stability and peace of mind.