For nearly 30 years, the so-called ‘4% rule’ has been a starting point for retirement planning conversations between financial planners and their clients. But as equity valuations such as the Shiller CAPE ratio have ratcheted up to nearly all-time highs in recent years, with bond yields simultaneously reaching all-time lows (suggesting below-average future returns in both asset classes), some experts have questioned whether a 4% initial withdrawal rate will continue to be ‘safe’ in the future. A new white paper by Morningstar feeds into this speculation, with its much-publicized conclusion stating that, given today’s market conditions, the future safe withdrawal rate should be lowered to 3.3%.
The Morningstar paper’s key insight is that expectations for future safe withdrawal rates should be adjusted based on current market conditions (which other research has supported). Accordingly, the paper’s authors use forward-looking return projections to calculate their future safe withdrawal rate estimates. But the investment return assumptions that Morningstar used for its analysis were so low – with real returns averaging just 5.7% for equities and 0.5% (!) for fixed income over 30 years – that, if those projections were to come to pass, the next 30 years would be among the very worst market environments in U.S. history.
While such conservative return estimates might make sense over a 10- to 20-year time horizon (since research has shown that CAPE ratios are strongly predictive of returns over that time range), extending those assumptions out to 30 years is arguably unrealistic. This is because there is no precedent – even in other eras with high equity valuations – for 30-year returns that low (other than the period spanning the late 19th and early 20th centuries, when financial panics and global war created a far more tumultuous and unpredictable environment than the comparatively stable world today). In reality, markets tend to revert to the mean, meaning that even the periods with the worst safe withdrawal rates in history contained intervals of offsetting below- and above-average returns, causing each to end out with near-average returns over the full 30-year horizon.
In this way, Morningstar’s choice to focus on (historically low) 30-year returns for its analysis disregards the evidence of what really drives safe withdrawal rates, which is the sequence of returns. Because the periods that have tested the 4% rule in the past were not necessarily those with the worst 30-year returns, but those whose returns in the first 10-15 years were so bad that retirees needed to withdraw too much of their portfolio to be able to recover once conditions improved. So in reality, Morningstar’s results may have been more realistic if they had only forecast 15-year instead of 30-year returns, since the 15-year period is both easier to predict on current market data and more predictive of safe withdrawal rates.
Ultimately, however, Morningstar’s conservative return assumptions – which are comparable to some of the worst periods in the past 140 years – actually serve to highlight the strength of the 4% rule, which was created to withstand just those types of worst-case scenarios. Which means that, even if their historically low projections do come to pass, resulting in returns equal to the worst return scenarios in history, a 4% initial withdrawal rate would still hold up. And while today’s market conditions do merit caution (as there is reason to believe that the next 15 years could experience below-average portfolio returns), in reality, such conditions were precisely what the 4% rule was created for, to begin with!
In 1994, financial planner William Bengen published his seminal research study on safe withdrawal rates. The paper established that, based on historical market data, a person who withdrew 4% of their portfolio’s value during their first year of retirement, then withdrew the same dollar amount adjusted for inflation in each subsequent year, would never run out of money by the end of a 30-year time horizon – even in the worst case sequence of returns ever experienced in the historical US data.
From this insight, the so-called “4% Rule” was born, and while it has been subject to numerous challenges and critiques over the years (with some calling it “too safe” and others claiming it is not safe enough), 4% remains anchored as at least a productive starting point for countless retirement planning conversations (before narrowing-in on more client-specific recommendations).
The most recent re-examination of the 4% rule, presented in a white paper from Morningstar titled “The State of Retirement Income: Safe Withdrawal Rates”, concludes that, based upon current market conditions including high equity valuations and low bond yields, the standard safe withdrawal rate should be lowered to 3.3%, which would amount to a relative 17.5% cut in lifetime retirement spending.
The paper garnered headlines based on this eyebrow-raising claim, but less-publicized (yet perhaps even more surprising) were the investment return assumptions that Morningstar used to reach its conclusion: 8.0% for equities and 2.7% for fixed income (translating to real returns of 5.7% for equities and 0.5% for fixed income, after factoring in the study’s assumed 2.21% inflation rate), which, when combined, would be among the worst portfolio returns ever achieved over a 30-year time horizon.
Bengen himself wrote a reaction and critique of the white paper following its publication, questioning Morningstar’s return assumptions and time horizon. But it is worth exploring in more depth what the Morningstar white paper actually says, and how its projections compare with the historical evidence that has traditionally been used for safe withdrawal rate research. Because even though today’s market conditions do create some risks for those who are retired (or who plan to do so soon), they are the same kinds of risks that the 4% rule was developed to defend against – meaning that, even if we were on the brink of an economic disaster on par with the Great Depression or the “stagflationary” 1970s, the 4% rule would still be sufficient to sustain withdrawals through retirement, as that’s what it was designed to defend against in the first place!
Dissecting Morningstar’s ‘State Of Retirement Income’ White Paper On Safe Withdrawal Rates
Understanding Safe Withdrawal Rates
Safe withdrawal rates begin with the assumption that a retiree will withdraw a certain percentage of their portfolio in their first year of retirement, then withdraw the same dollar amount (adjusted annually for inflation) in each subsequent year as a stable inflation-adjusted standard of living over a given time horizon (often assumed to be 30 years, though some research has modeled longer periods). A ‘safe’ withdrawal rate, then, is the highest initial withdrawal rate that will never result in the retiree fully drawing down their portfolio over the specified time horizon, even in the worst possible market conditions.
Safe withdrawal rate calculations, which usually assume a diversified mix of equity and fixed income investments, are generally based on three core factors:
- Equity returns;
- Fixed income returns; and
- The rate of inflation.
While equity and fixed income returns determine how much growth occurs in a diversified portfolio, the inflation rate determines the amount that is withdrawn in each subsequent year after the initial withdrawal. In general, then, higher stock and bond returns increase the safe withdrawal rate over a given time period (since more portfolio growth can sustain higher withdrawals), while higher inflation causes it to decrease (because higher inflation rates necessitate greater withdrawals and can deplete the portfolio sooner).
Just as important as the returns themselves is the order in which they are experienced. Two different retirement horizons may have the same average returns over 30 years, but depending on how those returns are sequenced, the safe withdrawal rates for one could be vastly different from the other. If a retiree is forced to withdraw too much of their portfolio in the early part of retirement – whether because of poor returns, high inflation, or a combination of both – it can be difficult for the portfolio’s value to recover later on, no matter how high the returns may be in the latter part of retirement.
Viewed another way, the sustainability of portfolio withdrawals is heavily influenced by the ability of the portfolio to grow at a pace above and beyond the rate of inflation, which helps to ameliorate the impact of an adverse early sequence if returns that recover (and then some) in later years.
Historical Safe Withdrawal Rate Research Versus Forward-Looking Estimates
Traditionally, safe withdrawal research has been based on historical data on equity and fixed income returns and inflation rates. Bengen’s original research to devise the 4% safe withdrawal rate was based on Ibbotson’s Stocks, Bonds, Bills, and Inflation (SBBI) data going back to 1926, while subsequent studies using Robert Shiller’s historical market research has increased the scope of the data set as far back as 1871. The key assumption, in either case, is that this historical data set – which now includes up to 122 ‘rolling’ 30-year time periods – will provide a reliable indicator of how bad future markets could behave, such that a withdrawal rate that was ‘safe’ in even the worst of those historical sequences using past market data would prove equally safe for future retirees that had similarly bad worst-case scenarios.
And in the 28 years since Bengen’s original paper (which itself nearly comprises an entire 30-year retirement time horizon), that assumption – and the 4% rule – has continued to hold up, even through the 2000 tech-bubble crash and 2008 global financial crisis (which, as bad as they were, still resulted in return sequences – so far – nowhere near bad enough to ‘break’ the 4% rule).
More recent research, however, has gone beyond using ‘purely’ historical data, creating forward-looking projections to estimate future safe withdrawal rates. This has been enabled by research showing that future returns can (to an extent) be predicted by current market conditions (e.g., P/E ratios for the stock market in the aggregate, and current bond yields) – thereby enabling researchers to project future long-term returns for equities and fixed income – such that future safe withdrawal rates themselves can also be predicted based on the current market environment.
The end result is the ability to create a forward-looking safe withdrawal rate that, rather than being simply based on the ‘worst-case’ scenario of all historical retirement scenarios, is adjusted for the current market conditions of someone currently reaching retirement. This forward-looking approach is the one that Morningstar’s study takes towards estimating future safe withdrawal rates.
How Morningstar’s Forward-Looking Return Assumptions Differ From Historical Averages
The basic concept of Morningstar’s paper is that today’s market conditions should be considered when planning for future retirement withdrawals. And there is no doubt that we are in an unusual market environment today. Despite the initial shock and ongoing uncertainty of the COVID-19 era, equities have roared back in the nearly two years since the start of the pandemic and are currently at their highest valuations (as measured by P/E10) since the dotcom bubble of the late 1990s. Meanwhile, interest rates (and bond yields in turn) have continued their steady march downward since the early 1980s.
The result, as shown in the chart below, is that we are currently experiencing among the highest-ever equity valuations, and the lowest-paying bonds, in 140 years of observable data… at the same time.
As many experts have warned in recent years, these conditions of high equity valuations and low bond yields suggest that future portfolio returns will be lower than what investors have recently experienced, as both numbers tend to (eventually) revert towards their historical averages.
Accordingly, the Morningstar study (which uses Morningstar Investment Services’ proprietary return forecasts as inputs for its return assumptions) projects well below-average returns for equities and fixed income compared to their historical averages:
As a result, all of the portfolio allocations used in Morningstar’s forward-looking analysis underperform the same portfolios using historical return data by upwards of 200 basis points per year when adjusted for inflation:
Most significantly, these low returns are projected for the entire 30 years of the study, leaving little wonder as to why the paper concludes that the safe withdrawal rate will also be lower than previous estimates that were based on historical returns.
Putting Morningstar’s Projections In Historical Context
Even though the point of Morningstar’s study is to use forward-looking estimates to calculate safe withdrawal rates, it’s worth examining their projections in a historical context to understand just how unusual of a low-return environment the study is predicting in the coming decades.
According to Robert Shiller’s historical market data, the average real return of a 60/40 portfolio between 1871 and 2021 was 6.01%. Over shorter time periods, the actual range of returns experienced may fluctuate widely around that average; as the period gets longer, however, the range of returns narrows on account of the tendency of markets to revert to the mean. In the chart below, the top and bottom quartiles of historical real returns for a 60/40 portfolio over rolling periods from 1 to 30 years are compared with both the average historical return and Morningstar’s projected average return for that portfolio:
As the chart shows, Morningstar’s projected average of a 3.47% real return for a 60/40 portfolio would not be unthinkable over a 1-year, 5-year, or even 10-year period (where a 3.47% average return would be below average but above, or nearly equal to, the bottom quartile of historical returns for those periods). But a 3.47% real return over 30 years would be a historical outlier, falling well below the 30-year bottom-quartile return of 4.76%. And in fact, the only era in the 140 years of observable data that has seen worse real returns over a 30-year period began over 130 years ago, in the early 1890s!
For context, that era in history included two widespread financial panics (in 1893 and 1907), a severe economic depression (from 1893 to 1897), the assassination of President William McKinley in 1901, the Spanish Flu epidemic, and the entirety of World War I. It also took place when the U.S. was on the Gold Standard and (up until 1914, when the U.S. Federal Reserve System was established) lacked a central bank with the tools to stabilize the financial system when crises arose, which led to more (and more severe) financial panics than the comparatively more stable Federal Reserve era.
In other words, if the next 30 years really do match Morningstar’s projections, that period will be comparable to the worst ever experienced in terms of real growth, which were among the most tumultuous times in the country’s history!
Furthermore, it’s important to note that the Morningstar study used Monte Carlo analyses to calculate their safe withdrawal rates, meaning that their ‘projected’ return is really the average of a distribution of returns used to run each scenario. So of all the simulated scenarios they used to run their Monte Carlo analysis, half were actually worse than that average (which was already low enough to be a historical outlier itself). And even that may understate how pessimistic the numbers in the analysis may have been, because (compared with historical data) Monte Carlo analysis can actually overstate the risk of extreme outlier scenarios if it fails to take into account the long-term tendency for markets to mean-revert (which it does not appear the Morningstar study considered).
Given the extreme pessimism of the projections used, it’s not surprising that the resulting safe withdrawal rate in Morningstar’s analysis was below 4%; in fact, arguably the bigger surprise is that it ‘only’ decreased by 0.7%, from 4% to 3.3%, given that more than 50% of their simulated scenarios were worse than almost every 30-year period of returns ever seen in US history!
Sequence Of Returns Matters More Than Average Returns For Safe Withdrawal Rates
Aside from their historically low return expectations, the methodology of Morningstar’s study also merits further examination. Morningstar’s method of calculating forward-looking safe withdrawal rates involved first estimating a 30-year average return and standard deviation value for a given investment portfolio, then using those inputs to run a Monte Carlo analysis for the 30-year period that simulated the amount and sequence of returns for each scenario (while setting a constant rate of inflation to determine portfolio withdrawals over the 30-year time horizon). The study had a 90% required success rate for its calculations, meaning that any initial withdrawal rate in which at least 90% of the Monte Carlo scenarios ended with a positive portfolio balance was considered safe.
The problem with this method, however, is that Monte Carlo analysis generates a random sequence of returns for each simulation, when, in reality, market returns are generally not random from one year to the next. Instead, markets have historically followed secular bull and bear market cycles, typically lasting 10-to-20 years each, during which average returns performed higher (in bull market cycles) or lower (in bear market cycles) than their overall historical averages. And because these market cycles often last for 10-to-20 years, the typical 30-year retirement horizon will likely encompass multiple cycles, including both bull and bear markets (which offset each other over the full time horizon, resulting in 30-year average returns that tend to fall close to the historical average).
For retirees, therefore, it is not a matter of whether they will experience a period of below-average returns during retirement – for such periods will occur in almost any retirement horizon – but where in the succession of bull and bear market cycles their retirement date takes place that will have the most impact on their safe withdrawal rate. Someone unfortunate enough to retire at the beginning of a bear market cycle may experience poor returns over the first 15 years of retirement – a costly sequence of returns that could result in a far lower safe withdrawal rate than if they had happened to retire near the start of a bull market cycle. But, because the bear market cycle would likely be followed by a bull market cycle which would offset the earlier poor returns, the retiree’s returns over the full 30-year period might end out close to the historical average (though they might take little comfort in that fact if they were forced to withdraw too much of their portfolio in the poorer-performing first half of retirement to benefit from the above-average returns in the second half!).
Historical evidence supports the idea that the worst safe withdrawal rates tend to occur when retirement coincides with the beginning of a bear market cycle. As shown below, the retirement horizons with the worst withdrawal rates in observable history all began with extended periods of flat-to-negative real returns. However, the second half of the retirement horizon (Years 15 – 30 after retirement) tells a completely different story, with markets shifting back to a bull cycle and turning sharply positive for the remainder.
And yet, even in these worst-case historical scenarios, the 4% rule still held – indeed, this is the very reason that 4% became known as the “safe” withdrawal rate to begin with. Meaning that, even if we are currently on the cusp of an extended bear market, the sequence of returns for the next 30 years would need to be worse not only than the historical average, but than all of the four scenarios above in order to ‘break’ the 4% rule.
Making Sustainable Withdrawal Rate Recommendations In The Current Low-Return Environment
What Today’s Valuations Really Tell Us About Future Return Assumptions
While there are concerns about the plausibility of Morningstar’s 30-year return assumptions and their relevance in calculating their safe withdrawal rate analysis, the basic premise of the paper – that current market conditions can be useful for estimating future safe withdrawal rates – still stands. However, it is important to understand what today’s valuations can actually tell us about future returns, and how that knowledge can be translated into future safe withdrawal rates, to make recommendations for sustainable withdrawals for people approaching retirement.
Prior research has shown that current equity valuations, in the form of the Shiller CAPE ratio, are strongly predictive of real returns in equities over a 10- to 20-year time horizon. However, they tend to lose that predictive power over longer time periods: Eventually, after market prices move in one direction for long enough to correct for the initial high (or low) valuations, the pendulum will reach the opposite extreme and prices will start to move in the other direction. Likewise, because it often takes a decade or more for this shift to occur, the CAPE ratio is similarly a poor predictor of market returns over less-than-10-year time periods as well.
Fortunately, real returns during the first 15 years of retirement (which the CAPE ratio predicts most effectively) are themselves predictive of the safe withdrawal rate for the entire 30-year retirement horizon, because those first 15 years are where the sequence of return risk for retirees is highest. So by extension, the CAPE ratio at the time of retirement can itself be a good predictor of the safe withdrawal rate over the next 30 years (which previous research has also supported).
Ultimately, though, this highlights a significant potential flaw in Morningstar’s return assumptions: Rather than use current market values to project future returns over the next 10-to-20 years (where current valuations have their greatest predictive power), and allowing them to subsequently recover (consistent with long-term mean reversion tendencies of markets), Morningstar extended those reduced projections over the entire 30-year horizon of the study, giving them results which look much more out of place in the historical context.
Because again, what today’s high valuations imply is not necessarily that we should expect portfolio returns to be significantly lower over the next 30 years (as the Morningstar study assumes). Rather, the historical evidence suggests that while real returns over the next 10-to-20 years are likely to fall below average (with the risk that, in extreme scenarios, inflation could push real returns near or below zero), 30-year real returns are likely to remain within about 100 basis points above or below the historical average, no matter what the results of the first 20 years looked like!
How Future Return Assumptions Affect Decision-Making Around Retirement
The distinction between Morningstar’s conservative expected return assumptions and the historical evidence is important, because each perspective uniquely affects the way financial advisors and their clients approach retirement planning. Assuming that returns will be low for the entire 30-year time horizon can lead to different decisions than assuming they will be low for ‘just’ the first 15-20 years before turning upward (i.e., mean-reverting) for the remainder.
For example, if returns were expected to be dramatically lower for the next 30 years, annuities might become a more attractive option for retirees to provide ‘guaranteed’ income and a hedge against the risk of outliving their assets (with the caveat that an annuity might not be quite so safe if returns were to end out being so low over the long term, since annuity companies rely on those returns to fund their contracts, too!). But if 30-year returns remained closer to the historical average (as has been the result when starting in other high-valuation periods in history, allowing 30 years for markets to have below-average returns and then recover), the decision could result in forgoing significant portfolio upside in exchange for the ‘guaranteed’ safety of lifetime annuity income.
Furthermore, the larger implications of such low returns over the next 30 years would be significant. Real returns of 3.5% for a 60/40 portfolio over a 30-year period have not been experienced since the 1800s; a return to that environment would imply not just a cyclical slowing of growth, but a major upheaval within our financial and economic system (and possibly society itself). Meaning that planning for retirement in a 3.5% real return world involves considering the other factors – climate change, economic inequality, or further pandemics, to name a few possibilities – that could lead to such a world, and how those could affect one’s lifestyle in retirement. But ultimately, while those issues are certainly salient for many today, their future effects are arguably not nearly certain enough to incorporate them into our assumptions for retirement planning as a baseline for retirement spending recommendations.
The key point, though, is that short of literally unprecedented changes to the US economic system, these patterns of reversion to the mean are what have given the 4% rule its staying power over the nearly three decades since it was developed. The rule was created to survive the worst possible sequence of returns over the typical retiree’s time horizon, so (by definition) any scenario that ‘breaks’ the 4% rule would need to be worse than every other 30-year period in the last 140 years for which we have available market data.
And simply put, there is nothing indicating that we are currently poised to enter such a period. Because even if today’s equity valuations and bond yields translate into low portfolio returns in the near future, and higher-than-expected inflation causes real returns to go flat or negative during that time, the data show that the 4% rule has already survived such scenarios – and would be expected to do so again, provided that the pattern of mean reversion continues to hold.
Safe Withdrawal Rates Are Still Just A Floor, Not A Ceiling
Because the 4% rule was created to survive the worst possible return environments for retirees, the vast majority of actual 30-year time periods in the historical data have supported a higher initial withdrawal rate than 4% (and often significantly higher). In other words, 4% can be considered a floor for retirement spending, not a ceiling, because anything less than a 4% initial withdrawal rate would virtually guarantee there would be excess money left ‘on the table’ after 30 years.
In fact, retirees over the last 140 years who strictly followed the 4% rule would have had only a 10% chance of finishing with anything less than their initial portfolio value… and an equally likely chance of finishing with more than six times their starting principal. So while planning for the downside risk of running out of money during retirement is often the primary concern of financial planners and their clients, it is also important to consider the ‘upside’ sequence-of-return risk that could result in large amounts of unspent money remaining at the end of retirement.
Fortunately, safe-withdrawal-rate research in more recent years has revealed strategies that can help retirees increase their initial withdrawal rates (and enjoy a higher level of income during their early, more active retirement years) while managing the downside of sequence-of-return risk, allowing retirees to use up more of their retirement savings without sacrificing the protection of their principal.
For example, for retirees who are willing to be somewhat flexible in their spending, dynamic flexible spending adjustments – e.g., making small (but permanent) spending cuts after years with negative portfolio returns – can increase the initial safe withdrawal rate by around 0.5%. Alternatively, a guardrails-based approach of maintaining spending within certain levels calculated yearly (using formulas such as portfolio withdrawal rates, Monte Carlo probabilities of success, or holistic risk-based factors) can be used to achieve sustainable withdrawals that are more specific to the individual retiree than a catchall method like the 4% rule.
Whatever the method of determining a safe withdrawal rate, though, it is crucial to understand the assumptions that go into the projections used to do so and how they relate to the historical evidence. While it is true that past performance does not guarantee future results, and that the absence of “Black Swans” in our historical evidence does not necessarily prove that they don’t exist, the reality is that we almost always rely (at least in part) on past data to make informed projections for the future.
However, Morningstar’s approach toward re-examining the 4% rule – projecting safe withdrawal rates based on drastically reduced 30-year return projections – ignores the way markets have actually behaved in the past, focusing only on reduced 30-year returns (which usually encompass multiple up-and-down market cycles) rather than on the crucial first 10 to 15 years of the retirement horizon when sequence of return risk is most relevant. In doing this, it appears that Morningstar’s study overestimates the likelihood of a potentially catastrophic decline in economic growth over the next 30 years, while ironically still underestimating the likelihood that the next 10 to 15 years could see flat to negative real returns in the event of higher-than-expected inflation (creating more risk for retirees and requiring careful monitoring to ensure the sustainability of their portfolio withdrawals). Which, fortunately, more dynamic withdrawal frameworks are built to accommodate anyway.
Ultimately, the key point is that the 4% rule has held up in some incredibly bleak periods in financial history, to such an extent that predicting its future demise is literally a prediction of a future that will be worse than anything ever seen in US history, including multiple financial crises and two global wars. The combination of factors that would need to occur simultaneously to ‘break’ the 4% rule – very low equity and bond returns, plus above-average inflation – has, in the past, rarely occurred, and the natural state of subsequent economic growth (that eventually lifts markets again), combined with real-world mechanics that impact the market (e.g., the ability for the Federal Reserve to increase interest rates to reduce inflation, or the ability for businesses to boost prices during inflationary periods, leading to higher earnings and increasing equity returns) may indeed prevent even as-bad-as-historical scenarios from ever happening (not to mention something worse). Which helps to explain why even retirees that began on the eve of the 2000 tech crash or the 2008 financial crisis are still on a ‘safe’ withdrawal rate trajectory today.
So for now, while today’s equity valuations and bond yields suggest that caution is merited when making withdrawal recommendations for retirees… this is not necessarily the environment that is going to ‘break’ the 4% rule; instead, it’s precisely what the 4% rule was made for in the first place!