In the past few years, the 4% rule has been challenged by those who claim its premise of 4% inflation-adjusted withdrawals is too optimistic under today’s market conditions. Others assert that more sophisticated approaches will yield better-than-4% results. I'll evaluate two alternatives – economic utility maximization and required minimum distributions (RMDs) – and also discuss the practical implications for advisors.
Advisors face a confusing array of choices in deciding how to advise clients about using savings to generate retirement income. In this blog post, retirement researcher Wade Pfau developed a chart displaying 34 different strategies for managing retirement income, and the list continues to grow. Evaluating all these choices would require a Piketty-length book.
I'll compare the 4% rule to two other approaches that have particular relevance for advisors. I'll also discuss variations on these strategies.
The 4% rule
The 4% rule was developed by financial planner William Bengen and described in his article “Determining Withdrawal Rates Using Historical Data” in the October 1994 issue of the Journal of Financial Planning. This is arguably the most famous article ever written about financial planning. Bengen's primary goal was to demonstrate that a safe retirement withdrawal rate was significantly less than one might assume based on historical average returns for stocks and bonds. He based the analysis on a withdrawal rate established at retirement as a percentage of savings, with the dollar amount of withdrawals increasing each year by the inflation rate.
In recent years, the 4% rule has been criticized as too aggressive given current low bond yields and lower return prospects for stocks, such as in this article by Michael Finke, Wade Pfau and David Blanchett. The rule has also been criticized as too inflexible, unaware of individual risk preferences and incapable of adapting to realized investment performance. Examples of such criticisms are “The 4% Rule – At What Price?” by Jason Scott, William Sharpe and John Watson and “Spending Retirement on Planet Vulcan: The Impact of Longevity Risk Aversion on Optimal Withdrawal Rates“ by Moshe Milevsky and Huaxiong Huang.
Such criticism has given rise to the growing number of proposed alternatives.
The economic approach – utility maximization
The 4% rule offers a simple rule of thumb of inflation-adjusted withdrawals and a "set it and forget it" approach. At the opposite extreme are methods of generating retirement income that rely on mathematical optimization with frequent updates to retirement plans. The particular optimization method I am most familiar with is the AACalc software developed by Gordon Irlam. My recent Advisor Perspectives article provides an overview, and a more in-depth treatment can be found in this Journal of Retirement article (available to subscribers) that I co-authored with Irlam.
AACalc and other optimization software programs are based on the economic principle of maximizing the utility of lifetime consumption. With life-cycle theory, it's easy to get lost in the math and the economics jargon, but optimization is conceptually straightforward. The focus of optimization is consumption spending over the retirement period. There is a trade-off between being able to spend high amounts with a lot of year-to-year variability and spending less on average but experience less fluctuation. The variability mostly reflects the aggressiveness of the investment policy. More-aggressive investing leads to more-variable spending, and vice versa. There are also choices about how much to front-load retirement spending toward the early years, when one is more likely to enjoy the benefits.
The strength of the optimization approach as applied in the AACalc software is that it uses stochastic dynamic programming, a very powerful tool that develops a plan for both year-by-year retirement spending and determines the optimum glide path for asset allocations. It does this by creating a "retirement map," which prescribes an asset allocation and recommended spending amount for each age and level of remaining savings. An individual can plot one's way through retirement by following the map.
A disadvantage of optimization approaches is that they require inputs such as risk-aversion coefficients and personal discount rates, which are difficult for advisors and clients to estimate. A challenge for the future will be to develop user-friendly questionnaires to determine the appropriate economic parameters.
The only commercially available software package that uses optimization is ESPlanner, developed by Laurence Kotlikoff of Boston University (see here). It has been available for more than a decade but has never gained the popularity of non-optimizing packages like MoneyGuide Pro. Irlam's AACalc software is available for free for those who wish to test it out. Also advocating optimization is a recent paper from J.P. Morgan that demonstrates the advantages of dynamically adapting asset allocation and withdrawal rates to make the most of retirement assets.
Optimization not only provides a powerful tool that can be used for retirement income management, but it can also be used to evaluate and compare various rule-of-thumb approaches. I'll examine another retirement income strategy, the required minimum distribution (RMD) approach, and show how it stacks up against the 4% rule based on optimization analysis.
Required minimum distributions (RMDs)
For tax-deferred investments such as 401(k)s and traditional IRAs, tax laws require taking withdrawals upon reaching age 70.5. The IRS publishes tables that specify the minimum percentage of remaining tax-deferred savings that must be withdrawn at each age. The percentages are intended to spread the withdrawal of savings over one’s life expectancy.
Economists have noted that withdrawals under this RMD method closely match the withdrawals prescribed by their life-cycle optimization models. This has given rise to research that compares RMDs and 4% rule withdrawals using optimized withdrawals as a benchmark. The comparison measure is related to economic utility, specifically what is known as the "certainty equivalent" of retirement consumption – the annual amount of consumption that has the same total utility value as the typically variable stream of consumption spending produced by the RMD or 4% rule withdrawal methods.
Wei Sun and Anthony Webb of the Center for Retirement Research at Boston College provided an extensive analysis of the RMD method in this 2012 paper and later summarized their work and offered new insights in this issue brief. They extended the RMD table back to age 65 using the same life tables that the IRS used to determine the RMD percentages starting at age 70.5. They found that the RMD approach performed significantly better than the 4% rule.
Irlam and the J.P. Morgan study have also provided comparisons of the RMD and 4% rule approaches that show RMD performing best. Exhibit 12 of the J.P. Morgan study compared year-by-year consumption spending patterns, and it is clear that the RMD approach closely matches the optimization consumption path. This chart provides a comparison of certainty equivalents from the three studies.
Comparison of withdrawal methods
Certainty equivalent consumption spending
Method |
Sun/Webb |
Irlam |
J.P. Morgan |
4% Rule |
0.56 |
0.71 |
0.87 |
RMD |
0.77 |
0.95 |
0.98 |
Optimized |
1.00 |
1.00 |
1.00 |
Source: Author's calculations from cited articles
Here is an example of how to read the chart: Sun/Webb estimated that if optimized withdrawals produced $100,000 of certainty equivalent annual income, the 4% rule would produce $56,000.
The results from the three studies are quite different, reflecting differences in underlying assumptions and methodology, but all studies show a significant advantage for RMD over the 4% rule.
A problem with RMD and optimized methods is that the recommended annual changes in consumption may be too disruptive. For example, under both methods, a 10% decrease in portfolio values will result in roughly a 10% decrease in withdrawals. The impact on consumption spending will depend on the share of spending provided by guaranteed income sources such as Social Security, pensions or annuities. The more income that is derived from those sources, the less the impact will be on spending.
Improving the methods
Improvements to retirement-income strategies can be accomplished by altering inflexible approaches like the 4% rule to increase flexibility or reducing the volatility associated with strategies like RMD or optimization. There may be other enhancements that provide more downside protection or that change the timing of spending in a favorable way.
Jonathan Guyton has advocated making the 4% rule more flexible by incorporating spending adjustments to reflect underlying investment performance. In this 2006 Journal of Financial Planning article, Guyton and William Klinger demonstrated that they could safely raise the initial withdrawal rate to between 5.1% and 5.6% as long as they incorporated certain decision rules to mitigate years of subpar investment performance. An example of one of the rules is that withdrawals increase each year by the inflation rate, except no inflation increase is allowed in years when the portfolio return is negative. The measurement criteria they used were the withdrawal rate and probability of depleting funds during a 40-year period. A potential future research project would be to measure results based on certainty equivalent consumption spending and test them against optimization models, as was done for the RMD analysis.
In terms of making RMD or optimization methods less volatile, the most straightforward approach is to increase fixed-income allocations. One way to do this is to follow a safety-first approach and build a higher floor of guaranteed lifetime income than Social Security offers by using a portion of savings to purchase an annuity such as a single-premium immediate annuity (SPIA). The higher proportion of guaranteed lifetime income would lessen the percentage impact of withdrawals on total consumption spending. A further reduction in spending volatility could be accomplished by increasing the bond allocation in the savings not used for annuity purchase.
For those who find full optimization too much of a "black box" and are more comfortable with rules of thumb, there are ways to fine-tune strategies similar to Guyton's fixes to the 4% rule. For example, Sun and Webb found that by modifying RMD to add interest and dividends to the withdrawals and thereby front-load consumption spending, they could produce results that were closer to optimal.
Others have approached retirement-income strategy analysis by carefully examining and adjusting the economic-utility analysis that supports optimization strategies. Jason Scott and John Watson of Financial Engines adjusted utility functions to recognize that individuals likely have a stronger desire for sustainable (non-decreasing) consumption spending than standard utility analysis assumes. They then developed the "floor-leverage rule" described in this paper. They found that by investing 85% of a portfolio in either bond ladders or fixed annuities like SPIAs, they could invest the remainder in a 3X leveraged stock portfolio and produce a consumption spending stream that would only increase and never decrease. Tests showed this approach produces nearly optimal results across a range of client preferences for sustainable spending.
Conclusion
Managing retirement income creates a lot of confusion for advisors when determining how best to deal with real-world client issues. This is a rich area for future research. A key goal is finding an approach that produces an attractive average consumption level without too much year-to-year spending volatility.
Another consideration is whether to rely on a rule-of-thumb strategy or use a software package that provides full optimization. I expect we will see significant growth in optimization software, particularly from "robo-advisor" companies like Betterment and Wealthfront, as they attempt to move from offering allocation advice for accumulators to providing computerized retirement spending and investment advice.
We are seeing the beginning of a development effort that is going to get much bigger.
Joe Tomlinson, an actuary and financial planner, is managing director of Tomlinson Financial Planning, LLC in Greenville, Maine. His practice focuses on retirement planning. He also does research and writing on financial planning and investment topics.
Read more articles by Joe Tomlinson
